Skip to content

Conversation

@hunhoffe
Copy link
Collaborator

@hunhoffe hunhoffe commented Dec 5, 2025

Add a codeowners file

Added

  • Code owners file

Changed

Removed

PR Merge Checklist

  1. The PR is rebased on the latest devel commit and pointing to devel.
  2. Your PR has been reviewed and approved.
  3. All checks are passing.

@github-actions
Copy link
Contributor

github-actions bot commented Dec 5, 2025

📊 Test Results for Test Example Applications

c6586c8 (2025_12_05_23_16_27)

IRONCLAD

Tested on 2025_12_05_23_16_27 at commit c6586c8.

Test Checks TTFT (mean)TPS (mean)
llama_3.2_1b ✅ 5/5 4.39 4.65
📈 Trends (vs main branch) for Test Example Applications

c6586c8 (2025_12_05_23_16_27)

IRONCLAD Trends

llama_3.2_1b

Commit/Date Num Tokens (max)Num Tokens (mean)Num Tokens (median)Num Tokens (min)Num Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)Total (max)Total (mean)Total (median)Total (min)Total (stddev)
c6586c8 — 2025-12-05 23:11:4040.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)0.00 (n/a)4.75 (+0.85%)4.65 (+0.26%)4.63 (-0.22%)4.57 (+0.44%)0.07 (+27.17%)4.44 (+0.69%)4.39 (+0.05%)4.40 (+0.27%)4.33 (-0.92%)0.04 (+157.84%)12.94 (-0.18%)12.78 (-0.20%)12.80 (+0.04%)12.61 (-0.49%)0.14 (+51.10%)
130b6ea — 2025-12-05 21:33:1240.00 (n/a)40.00 (n/a)40.00 (n/a)40.00 (n/a)0.00 (n/a)4.71 (n/a)4.64 (n/a)4.64 (n/a)4.55 (n/a)0.05 (n/a)4.41 (n/a)4.39 (n/a)4.38 (n/a)4.37 (n/a)0.01 (n/a)12.96 (n/a)12.80 (n/a)12.80 (n/a)12.67 (n/a)0.09 (n/a)

@github-actions
Copy link
Contributor

github-actions bot commented Dec 5, 2025

📊 Test Results for Small Benchmark/Test Suite

c6586c8 (2025_12_05_23_49_47)

IRONCLAD

Tested on 2025_12_05_23_49_47 at commit c6586c8.

Test Checks Latency (mean)Bandwidth (mean)GFLOP/s (mean)
axpy_1_cols_2_channels_2048_tile_2048_3.0 ✅ 5/5 141.18 0.09 n/a
axpy_2_cols_2_channels_2048_tile_1024_3.0 ✅ 5/5 108.22 0.12 n/a
axpy_4_cols_2_channels_2048_tile_512_3.0 ✅ 5/5 119.06 0.11 n/a
axpy_8_cols_2_channels_2048_tile_256_3.0 ✅ 5/5 154.42 0.08 n/a
dequant_1_cols_1_channels_2048_tile_2048 ✅ 5/5 126.78 0.04 n/a
dequant_1_cols_2_channels_2048_tile_1024 ✅ 5/5 106.00 0.05 n/a
dequant_2_cols_1_channels_2048_tile_1024 ✅ 5/5 108.20 0.05 n/a
dequant_2_cols_2_channels_2048_tile_512 ✅ 5/5 113.88 0.05 n/a
dequant_4_cols_1_channels_2048_tile_512 ✅ 5/5 126.00 0.04 n/a
dequant_4_cols_2_channels_2048_tile_256 ✅ 5/5 130.12 0.04 n/a
dequant_8_cols_1_channels_2048_tile_256 ✅ 5/5 117.74 0.05 n/a
dequant_8_cols_2_channels_2048_tile_128 ✅ 5/5 140.76 0.04 n/a
eltwise_add_1_cols_2_channels_2048_tile_2048 ✅ 5/5 108.28 0.11 n/a
eltwise_add_2_cols_2_channels_2048_tile_1024 ✅ 5/5 120.32 0.11 n/a
eltwise_add_4_cols_2_channels_2048_tile_512 ✅ 5/5 112.82 0.11 n/a
eltwise_add_8_cols_2_channels_2048_tile_256 ✅ 5/5 147.08 0.09 n/a
eltwise_mul_1_cols_2_channels_2048_tile_2048 ✅ 5/5 122.40 0.10 n/a
eltwise_mul_2_cols_2_channels_2048_tile_1024 ✅ 5/5 140.00 0.09 n/a
eltwise_mul_4_cols_2_channels_2048_tile_512 ✅ 5/5 123.98 0.10 n/a
eltwise_mul_8_cols_2_channels_2048_tile_256 ✅ 5/5 136.32 0.09 n/a
gelu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 119.88 0.07 n/a
gelu_1_cols_2_channels_2048_tile_1024 ✅ 5/5 132.88 0.06 n/a
gelu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 117.88 0.07 n/a
gelu_2_cols_2_channels_2048_tile_512 ✅ 5/5 122.00 0.07 n/a
gelu_4_cols_1_channels_2048_tile_512 ✅ 5/5 111.94 0.07 n/a
gelu_4_cols_2_channels_2048_tile_256 ✅ 5/5 144.82 0.06 n/a
gelu_8_cols_1_channels_2048_tile_256 ✅ 5/5 131.10 0.07 n/a
gelu_8_cols_2_channels_2048_tile_128 ✅ 5/5 163.62 0.05 n/a
gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0 ✅ 5/5 24520.66 1.03 n/a
gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0 ✅ 5/5 25978.06 0.97 n/a
gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0 ✅ 5/5 27478.34 0.92 n/a
layer_norm_1_cols_1_channels_2048_tile_2048 ✅ 5/5 140.18 0.06 n/a
layer_norm_1_cols_2_channels_2048_tile_1024 ✅ 5/5 129.74 0.06 n/a
layer_norm_2_cols_1_channels_2048_tile_1024 ✅ 5/5 122.12 0.07 n/a
layer_norm_2_cols_2_channels_2048_tile_512 ✅ 5/5 143.24 0.06 n/a
layer_norm_4_cols_1_channels_2048_tile_512 ✅ 5/5 122.60 0.07 n/a
layer_norm_4_cols_2_channels_2048_tile_256 ✅ 5/5 141.30 0.06 n/a
layer_norm_8_cols_1_channels_2048_tile_256 ✅ 5/5 135.78 0.06 n/a
layer_norm_8_cols_2_channels_2048_tile_128 ✅ 5/5 156.98 0.05 n/a
matrix_vector_mul_128x128_32_1col ✅ 5/5 n/a 0.32 n/a
matrix_vector_mul_2048x8192_1_1col ✅ 5/5 n/a 13.06 n/a
matrix_vector_mul_2048x8192_1_2col ✅ 5/5 n/a 24.42 n/a
matrix_vector_mul_2048x8192_1_4col ✅ 5/5 n/a 39.85 n/a
matrix_vector_mul_2048x8192_1_8col ✅ 5/5 n/a 44.08 n/a
matrix_vector_mul_8192x2048_4_1col ✅ 5/5 n/a 12.95 n/a
matrix_vector_mul_8192x2048_4_2col ✅ 5/5 n/a 23.38 n/a
matrix_vector_mul_8192x2048_4_4col ✅ 5/5 n/a 39.03 n/a
matrix_vector_mul_8192x2048_4_8col ✅ 5/5 n/a 45.77 n/a
mem_copy_16_cores_2_chans_2048_tile_128_False ✅ 5/5 167.60 0.05 n/a
mem_copy_1_cores_1_chans_2048_tile_2048_False ✅ 5/5 147.64 0.06 n/a
mem_copy_2_cores_1_chans_2048_tile_1024_False ✅ 5/5 145.92 0.06 n/a
mem_copy_2_cores_2_chans_2048_tile_1024_False ✅ 5/5 137.12 0.06 n/a
mem_copy_4_cores_1_chans_2048_tile_512_False ✅ 5/5 134.78 0.06 n/a
mem_copy_4_cores_2_chans_2048_tile_512_False ✅ 5/5 120.42 0.07 n/a
mem_copy_8_cores_1_chans_2048_tile_256_False ✅ 5/5 136.36 0.06 n/a
mem_copy_8_cores_2_chans_2048_tile_256_False ✅ 5/5 132.24 0.06 n/a
mha ✅ 5/5 47230.52 0.18 n/a
relu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 120.32 0.07 n/a
relu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 131.82 0.06 n/a
relu_4_cols_1_channels_2048_tile_512 ✅ 5/5 117.80 0.07 n/a
relu_8_cols_1_channels_2048_tile_256 ✅ 5/5 143.28 0.06 n/a
rms_norm_1_cols_1_channels_2048_tile_2048 ✅ 5/5 139.06 0.06 n/a
rms_norm_1_cols_2_channels_2048_tile_1024 ✅ 5/5 103.34 0.08 n/a
rms_norm_2_cols_1_channels_2048_tile_1024 ✅ 5/5 126.88 0.07 n/a
rms_norm_2_cols_2_channels_2048_tile_512 ✅ 5/5 117.18 0.07 n/a
rms_norm_4_cols_1_channels_2048_tile_512 ✅ 5/5 134.86 0.06 n/a
rms_norm_4_cols_2_channels_2048_tile_256 ✅ 5/5 120.46 0.07 n/a
rms_norm_8_cols_1_channels_2048_tile_256 ✅ 5/5 140.60 0.06 n/a
rms_norm_8_cols_2_channels_2048_tile_128 ✅ 5/5 117.84 0.07 n/a
rope_1_cols_2_channels_4096_tile_4096_0 ✅ 5/5 146.76 0.17 n/a
rope_2_cols_2_channels_4096_tile_2048_0 ✅ 5/5 141.78 0.17 n/a
rope_4_cols_2_channels_4096_tile_1024_0 ✅ 5/5 144.02 0.18 n/a
rope_8_cols_2_channels_4096_tile_512_0 ✅ 5/5 163.28 0.16 n/a
sigmoid_1_cols_1_channels_2048_tile_2048 ✅ 5/5 111.62 0.08 n/a
sigmoid_2_cols_1_channels_2048_tile_1024 ✅ 5/5 119.88 0.07 n/a
sigmoid_4_cols_1_channels_2048_tile_512 ✅ 5/5 114.68 0.07 n/a
sigmoid_8_cols_1_channels_2048_tile_256 ✅ 5/5 128.52 0.07 n/a
silu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 132.60 0.06 n/a
silu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 122.20 0.07 n/a
silu_4_cols_1_channels_2048_tile_512 ✅ 5/5 126.86 0.07 n/a
silu_8_cols_1_channels_2048_tile_256 ✅ 5/5 140.02 0.06 n/a
softmax_1_cols_2_channels_4096_tile_2048 ✅ 5/5 142.02 0.12 n/a
softmax_2_cols_2_channels_4096_tile_1024 ✅ 5/5 115.66 0.14 n/a
softmax_2_cols_2_channels_4096_tile_512 ✅ 5/5 141.80 0.13 n/a
swiglu_decode_1x2048x2048 ✅ 5/5 4128.28 0.00 n/a
tanh_1_cols_1_channels_2048_tile_2048 ✅ 5/5 107.90 0.08 n/a
tanh_2_cols_1_channels_2048_tile_1024 ✅ 5/5 140.40 0.06 n/a
tanh_4_cols_1_channels_2048_tile_512 ✅ 5/5 132.70 0.06 n/a
tanh_8_cols_1_channels_2048_tile_256 ✅ 5/5 146.86 0.06 n/a
transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s ✅ 5/5 148.14 3.56 n/a
transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s ✅ 5/5 131.24 4.05 n/a
weighted_rms_norm_1_cols_2_channels_2048_weights_2048 ✅ 5/5 120.74 0.07 n/a
weighted_rms_norm_2_cols_2_channels_2048_weights_1024 ✅ 5/5 128.70 0.07 n/a
weighted_rms_norm_4_cols_2_channels_2048_weights_512 ✅ 5/5 130.02 0.06 n/a
weighted_rms_norm_8_cols_2_channels_2048_weights_256 ✅ 5/5 149.18 0.06 n/a
📈 Trends (vs main branch) for Small Benchmark/Test Suite

c6586c8 (2025_12_05_23_49_47)

IRONCLAD Trends

axpy_1_cols_2_channels_2048_tile_2048_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.11 (-12.72%)0.09 (-14.80%)0.09 (-18.30%)0.07 (-3.41%)0.02 (-16.30%)168.20 (+3.57%)141.18 (+16.72%)143.20 (+22.39%)108.10 (+14.51%)22.79 (-1.39%)
130b6ea — 2025-12-05 21:15:490.13 (n/a)0.11 (n/a)0.10 (n/a)0.08 (n/a)0.02 (n/a)162.40 (n/a)120.96 (n/a)117.00 (n/a)94.40 (n/a)23.11 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.14 (+33.87%)0.12 (+26.88%)0.12 (+34.90%)0.09 (+13.34%)0.02 (+77.40%)134.10 (-11.78%)108.22 (-20.11%)105.80 (-25.91%)85.90 (-25.30%)17.68 (+17.63%)
130b6ea — 2025-12-05 21:15:490.11 (n/a)0.09 (n/a)0.09 (n/a)0.08 (n/a)0.01 (n/a)152.00 (n/a)135.46 (n/a)142.80 (n/a)115.00 (n/a)15.03 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.14 (+16.85%)0.11 (+8.25%)0.10 (+1.85%)0.09 (+18.35%)0.02 (+17.99%)140.80 (-15.54%)119.06 (-7.55%)119.60 (-1.81%)88.90 (-14.44%)20.13 (-12.28%)
130b6ea — 2025-12-05 21:15:490.12 (n/a)0.10 (n/a)0.10 (n/a)0.07 (n/a)0.02 (n/a)166.70 (n/a)128.78 (n/a)121.80 (n/a)103.90 (n/a)22.95 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (-8.97%)0.08 (-1.80%)0.08 (-3.82%)0.07 (+0.49%)0.01 (-34.93%)183.50 (-0.49%)154.42 (+0.81%)149.80 (+3.96%)139.60 (+9.83%)15.25 (-28.25%)
130b6ea — 2025-12-05 21:15:490.10 (n/a)0.08 (n/a)0.09 (n/a)0.07 (n/a)0.01 (n/a)184.40 (n/a)153.18 (n/a)144.10 (n/a)127.10 (n/a)21.26 (n/a)

dequant_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.05 (+16.23%)0.04 (+13.93%)0.04 (+2.08%)0.03 (+0.62%)0.01 (+76.48%)159.40 (-0.62%)126.78 (-10.16%)133.60 (-2.05%)98.30 (-14.00%)24.51 (+41.10%)
130b6ea — 2025-12-05 21:15:490.05 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.00 (n/a)160.40 (n/a)141.12 (n/a)136.40 (n/a)114.30 (n/a)17.37 (n/a)

dequant_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.06 (-4.36%)0.05 (-1.59%)0.05 (-13.95%)0.05 (+14.99%)0.00 (-37.33%)116.20 (-13.02%)106.00 (-0.08%)110.60 (+16.30%)91.90 (+4.55%)10.22 (-42.90%)
130b6ea — 2025-12-05 21:15:490.06 (n/a)0.05 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)133.60 (n/a)106.08 (n/a)95.10 (n/a)87.90 (n/a)17.90 (n/a)

dequant_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.06 (-5.52%)0.05 (-7.94%)0.05 (-5.32%)0.04 (+10.69%)0.01 (-17.33%)129.50 (-9.69%)108.20 (+7.11%)103.10 (+5.64%)83.50 (+5.83%)17.94 (-20.48%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)143.40 (n/a)101.02 (n/a)97.60 (n/a)78.90 (n/a)22.56 (n/a)

dequant_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.05 (+6.29%)0.05 (+8.58%)0.05 (-6.40%)0.04 (+25.32%)0.00 (-39.95%)127.40 (-20.23%)113.88 (-9.85%)116.50 (+6.88%)100.50 (-5.90%)10.20 (-54.62%)
130b6ea — 2025-12-05 21:15:490.05 (n/a)0.04 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)159.70 (n/a)126.32 (n/a)109.00 (n/a)106.80 (n/a)22.48 (n/a)

dequant_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.06 (-3.22%)0.04 (-4.11%)0.04 (-7.74%)0.03 (-8.73%)0.01 (-5.60%)171.00 (+9.55%)126.00 (+4.25%)124.60 (+8.44%)91.20 (+3.28%)26.01 (+6.61%)
130b6ea — 2025-12-05 21:15:490.06 (n/a)0.05 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)156.10 (n/a)120.86 (n/a)114.90 (n/a)88.30 (n/a)24.40 (n/a)

dequant_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.05 (-2.61%)0.04 (+2.75%)0.04 (+0.91%)0.03 (+9.99%)0.00 (-29.73%)152.90 (-9.04%)130.12 (-4.14%)131.00 (-0.91%)110.30 (+2.70%)14.94 (-34.78%)
130b6ea — 2025-12-05 21:15:490.05 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)168.10 (n/a)135.74 (n/a)132.20 (n/a)107.40 (n/a)22.91 (n/a)

dequant_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.05 (+15.19%)0.05 (+19.71%)0.05 (+21.06%)0.03 (+10.53%)0.01 (+31.08%)154.20 (-9.51%)117.74 (-15.97%)112.20 (-17.38%)103.90 (-13.13%)18.67 (+4.10%)
130b6ea — 2025-12-05 21:15:490.04 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.00 (n/a)170.40 (n/a)140.12 (n/a)135.80 (n/a)119.60 (n/a)17.94 (n/a)

dequant_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.05 (+10.43%)0.04 (+14.12%)0.04 (+19.96%)0.03 (+7.86%)0.01 (+12.38%)183.70 (-7.32%)140.76 (-12.03%)138.50 (-16.67%)96.40 (-9.48%)29.34 (-1.75%)
130b6ea — 2025-12-05 21:15:490.05 (n/a)0.03 (n/a)0.03 (n/a)0.03 (n/a)0.01 (n/a)198.20 (n/a)160.00 (n/a)166.20 (n/a)106.50 (n/a)29.86 (n/a)

eltwise_add_1_cols_2_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.13 (-5.57%)0.11 (+13.10%)0.11 (+34.91%)0.11 (+33.71%)0.01 (-63.37%)116.10 (-25.19%)108.28 (-16.46%)113.60 (-25.90%)93.50 (+5.89%)8.63 (-71.69%)
130b6ea — 2025-12-05 21:15:490.14 (n/a)0.10 (n/a)0.08 (n/a)0.08 (n/a)0.03 (n/a)155.20 (n/a)129.62 (n/a)153.30 (n/a)88.30 (n/a)30.47 (n/a)

eltwise_add_2_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.14 (+18.13%)0.11 (+8.58%)0.12 (+14.60%)0.08 (+1.27%)0.02 (+46.17%)162.80 (-1.27%)120.32 (-5.94%)106.80 (-12.75%)90.40 (-15.28%)26.94 (+25.17%)
130b6ea — 2025-12-05 21:15:490.12 (n/a)0.10 (n/a)0.10 (n/a)0.07 (n/a)0.02 (n/a)164.90 (n/a)127.92 (n/a)122.40 (n/a)106.70 (n/a)21.52 (n/a)

eltwise_add_4_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.13 (+16.03%)0.11 (+34.03%)0.12 (+62.03%)0.09 (+31.48%)0.02 (-10.27%)139.90 (-23.97%)112.82 (-26.76%)102.90 (-38.31%)95.50 (-13.81%)16.74 (-41.63%)
130b6ea — 2025-12-05 21:15:490.11 (n/a)0.08 (n/a)0.07 (n/a)0.07 (n/a)0.02 (n/a)184.00 (n/a)154.04 (n/a)166.80 (n/a)110.80 (n/a)28.67 (n/a)

eltwise_add_8_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.13 (+15.78%)0.09 (+3.55%)0.08 (-2.51%)0.06 (-9.05%)0.03 (+100.30%)195.60 (+9.95%)147.08 (+2.75%)146.40 (+2.59%)97.10 (-13.61%)41.38 (+92.97%)
130b6ea — 2025-12-05 21:15:490.11 (n/a)0.09 (n/a)0.09 (n/a)0.07 (n/a)0.01 (n/a)177.90 (n/a)143.14 (n/a)142.70 (n/a)112.40 (n/a)21.44 (n/a)

eltwise_mul_1_cols_2_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.13 (+15.41%)0.10 (+11.45%)0.11 (+16.60%)0.08 (+23.63%)0.02 (+11.85%)149.60 (-19.14%)122.40 (-10.63%)113.80 (-14.24%)95.80 (-13.38%)19.78 (-22.64%)
130b6ea — 2025-12-05 21:15:490.11 (n/a)0.09 (n/a)0.09 (n/a)0.07 (n/a)0.01 (n/a)185.00 (n/a)136.96 (n/a)132.70 (n/a)110.60 (n/a)25.57 (n/a)

eltwise_mul_2_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.13 (+3.10%)0.09 (-15.36%)0.08 (-20.77%)0.07 (-25.63%)0.02 (+120.86%)164.30 (+34.45%)140.00 (+21.55%)146.80 (+26.23%)96.40 (-3.02%)22.96 (+176.18%)
130b6ea — 2025-12-05 21:15:490.12 (n/a)0.11 (n/a)0.11 (n/a)0.10 (n/a)0.01 (n/a)122.20 (n/a)115.18 (n/a)116.30 (n/a)99.40 (n/a)8.31 (n/a)

eltwise_mul_4_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.11 (-8.53%)0.10 (-8.98%)0.11 (-12.05%)0.08 (+3.17%)0.01 (-20.45%)145.60 (-3.06%)123.98 (+9.00%)115.30 (+13.71%)108.60 (+9.37%)16.17 (-15.93%)
130b6ea — 2025-12-05 21:15:490.12 (n/a)0.11 (n/a)0.12 (n/a)0.08 (n/a)0.02 (n/a)150.20 (n/a)113.74 (n/a)101.40 (n/a)99.30 (n/a)19.23 (n/a)

eltwise_mul_8_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.11 (-2.14%)0.09 (+11.29%)0.09 (+5.01%)0.07 (+34.67%)0.01 (-28.24%)164.80 (-25.73%)136.32 (-12.69%)141.40 (-4.78%)115.90 (+2.20%)18.28 (-48.83%)
130b6ea — 2025-12-05 21:15:490.11 (n/a)0.08 (n/a)0.08 (n/a)0.06 (n/a)0.02 (n/a)221.90 (n/a)156.14 (n/a)148.50 (n/a)113.40 (n/a)35.72 (n/a)

gelu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (-1.35%)0.07 (+7.16%)0.07 (+18.15%)0.06 (-2.87%)0.01 (-7.09%)147.20 (+2.94%)119.88 (-6.84%)115.50 (-15.32%)101.20 (+1.40%)15.89 (-2.41%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.06 (n/a)0.06 (n/a)0.06 (n/a)0.01 (n/a)143.00 (n/a)128.68 (n/a)136.40 (n/a)99.80 (n/a)16.29 (n/a)

gelu_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (-3.09%)0.06 (-9.48%)0.06 (-15.30%)0.05 (-11.35%)0.01 (+58.42%)151.80 (+12.78%)132.88 (+12.08%)141.80 (+18.07%)107.50 (+3.17%)19.15 (+85.53%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)134.60 (n/a)118.56 (n/a)120.10 (n/a)104.20 (n/a)10.32 (n/a)

gelu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (-16.10%)0.07 (-5.24%)0.07 (-14.02%)0.06 (+10.89%)0.01 (-63.00%)133.70 (-9.84%)117.88 (+1.83%)117.70 (+16.30%)107.30 (+19.22%)9.37 (-61.34%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.07 (n/a)0.08 (n/a)0.06 (n/a)0.01 (n/a)148.30 (n/a)115.76 (n/a)101.20 (n/a)90.00 (n/a)24.24 (n/a)

gelu_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+1.79%)0.07 (-3.30%)0.07 (-8.95%)0.06 (-0.02%)0.01 (-4.33%)133.60 (+0.00%)122.00 (+3.34%)125.30 (+9.82%)105.40 (-1.77%)10.16 (-5.60%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)133.60 (n/a)118.06 (n/a)114.10 (n/a)107.30 (n/a)10.76 (n/a)

gelu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (-2.46%)0.07 (+18.89%)0.07 (+18.14%)0.06 (+36.42%)0.01 (-39.27%)128.80 (-26.73%)111.94 (-18.84%)110.80 (-15.36%)93.80 (+2.51%)13.51 (-53.40%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)175.80 (n/a)137.92 (n/a)130.90 (n/a)91.50 (n/a)29.00 (n/a)

gelu_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (+5.83%)0.06 (+3.53%)0.06 (+8.06%)0.05 (-4.44%)0.01 (+30.88%)172.70 (+4.67%)144.82 (-2.79%)144.30 (-7.44%)121.20 (-5.53%)18.38 (+29.53%)
130b6ea — 2025-12-05 21:15:490.06 (n/a)0.06 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)165.00 (n/a)148.98 (n/a)155.90 (n/a)128.30 (n/a)14.19 (n/a)

gelu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+6.24%)0.07 (-0.88%)0.07 (+8.64%)0.04 (-22.32%)0.01 (+74.92%)182.30 (+28.74%)131.10 (+4.76%)115.80 (-7.95%)97.60 (-5.88%)31.24 (+110.15%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)141.60 (n/a)125.14 (n/a)125.80 (n/a)103.70 (n/a)14.87 (n/a)

gelu_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.06 (-20.01%)0.05 (-4.41%)0.05 (+1.30%)0.04 (-1.85%)0.01 (-48.99%)200.40 (+1.88%)163.62 (+2.28%)159.50 (-1.30%)145.90 (+25.02%)19.08 (-33.41%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)196.70 (n/a)159.98 (n/a)161.60 (n/a)116.70 (n/a)28.66 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.72 (-0.11%)3.55 (-2.75%)3.56 (-2.75%)3.36 (-5.33%)0.12 (+76.10%)2725.23 (-0.11%)2602.52 (-2.75%)2612.11 (-2.75%)2459.89 (-5.33%)86.27 (+76.10%)6984.00 (+5.63%)6608.60 (+2.91%)6577.00 (+2.83%)6304.00 (+0.11%)221.87 (+87.15%)
9e98049 — 2025-11-25 20:50:553.72 (n/a)3.65 (n/a)3.66 (n/a)3.54 (n/a)0.07 (n/a)2728.26 (n/a)2676.06 (n/a)2686.03 (n/a)2598.29 (n/a)48.99 (n/a)6612.00 (n/a)6422.00 (n/a)6396.00 (n/a)6297.00 (n/a)118.55 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.40 (-3.14%)3.38 (-0.82%)3.38 (-0.82%)3.35 (+4.93%)0.02 (-84.51%)2493.81 (-3.14%)2475.27 (-0.82%)2479.06 (-0.82%)2453.92 (+4.93%)13.33 (-84.51%)7001.00 (-4.70%)6940.80 (+0.71%)6930.00 (+0.83%)6889.00 (+3.24%)37.43 (-84.76%)
9e98049 — 2025-11-25 20:50:553.51 (n/a)3.40 (n/a)3.41 (n/a)3.19 (n/a)0.12 (n/a)2574.53 (n/a)2495.72 (n/a)2499.62 (n/a)2338.67 (n/a)86.02 (n/a)7346.00 (n/a)6892.20 (n/a)6873.00 (n/a)6673.00 (n/a)245.57 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.02 (+0.52%)2.87 (-2.05%)2.86 (-2.24%)2.70 (-6.16%)0.11 (+131.66%)2216.19 (+0.52%)2104.44 (-2.05%)2094.08 (-2.24%)1980.84 (-6.16%)80.25 (+131.67%)8673.00 (+6.56%)8175.60 (+2.21%)8204.00 (+2.29%)7752.00 (-0.51%)314.03 (+145.36%)
9e98049 — 2025-11-25 20:50:553.01 (n/a)2.93 (n/a)2.92 (n/a)2.88 (n/a)0.05 (n/a)2204.81 (n/a)2148.41 (n/a)2142.13 (n/a)2110.81 (n/a)34.64 (n/a)8139.00 (n/a)7998.60 (n/a)8020.00 (n/a)7792.00 (n/a)127.99 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:401.04 (+0.58%)1.03 (-0.13%)1.03 (-0.06%)1.01 (-1.17%)0.01 (+110.51%)24867.90 (+1.19%)24520.66 (+0.14%)24545.60 (+0.06%)24126.10 (-0.58%)235.65 (+111.81%)
130b6ea — 2025-12-05 21:15:491.04 (n/a)1.03 (n/a)1.03 (n/a)1.02 (n/a)0.00 (n/a)24576.00 (n/a)24485.82 (n/a)24530.40 (n/a)24266.40 (n/a)111.26 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.97 (-1.74%)0.97 (-0.54%)0.97 (-0.02%)0.96 (-0.36%)0.00 (-58.34%)26120.10 (+0.36%)25978.06 (+0.54%)25921.70 (+0.02%)25896.40 (+1.77%)86.14 (-57.37%)
130b6ea — 2025-12-05 21:15:490.99 (n/a)0.97 (n/a)0.97 (n/a)0.97 (n/a)0.01 (n/a)26025.30 (n/a)25839.34 (n/a)25916.70 (n/a)25446.00 (n/a)202.05 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.92 (-0.30%)0.92 (+0.00%)0.92 (-0.43%)0.91 (+0.59%)0.00 (-66.40%)27588.40 (-0.59%)27478.34 (-0.01%)27473.50 (+0.43%)27402.40 (+0.30%)61.28 (-66.46%)
130b6ea — 2025-12-05 21:15:490.92 (n/a)0.92 (n/a)0.92 (n/a)0.91 (n/a)0.01 (n/a)27751.40 (n/a)27480.74 (n/a)27354.80 (n/a)27320.50 (n/a)182.69 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
31ab24e — 2025-11-22 02:01:563.65 (-1.06%)3.59 (+0.83%)3.63 (-0.26%)3.45 (+2.66%)0.07 (-44.71%)2678.08 (-1.06%)2632.13 (+0.83%)2658.60 (-0.26%)2528.31 (+2.66%)53.86 (-44.71%)6795.00 (-2.59%)6529.80 (-0.92%)6462.00 (+0.26%)6415.00 (+1.07%)137.14 (-45.27%)
d9a64c5 — 2025-11-21 20:49:353.69 (n/a)3.56 (n/a)3.64 (n/a)3.36 (n/a)0.13 (n/a)2706.77 (n/a)2610.50 (n/a)2665.61 (n/a)2462.71 (n/a)97.41 (n/a)6976.00 (n/a)6590.40 (n/a)6445.00 (n/a)6347.00 (n/a)250.58 (n/a)

layer_norm_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.06 (-32.60%)0.06 (-28.50%)0.06 (-33.90%)0.06 (+6.73%)0.00 (-78.74%)148.80 (-6.30%)140.18 (+33.38%)141.90 (+51.28%)126.10 (+48.35%)7.55 (-72.19%)
130b6ea — 2025-12-05 21:15:490.10 (n/a)0.08 (n/a)0.09 (n/a)0.05 (n/a)0.02 (n/a)158.80 (n/a)105.10 (n/a)93.80 (n/a)85.00 (n/a)27.15 (n/a)

layer_norm_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-5.67%)0.06 (-0.47%)0.06 (-1.02%)0.05 (-0.61%)0.01 (-5.32%)152.00 (+0.60%)129.74 (+0.40%)131.50 (+1.08%)113.90 (+5.95%)13.92 (+0.40%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)151.10 (n/a)129.22 (n/a)130.10 (n/a)107.50 (n/a)13.87 (n/a)

layer_norm_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (+25.96%)0.07 (+15.08%)0.08 (+24.80%)0.04 (-13.59%)0.01 (+144.25%)185.80 (+15.69%)122.12 (-8.92%)105.90 (-19.83%)96.00 (-20.60%)33.10 (+129.41%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)160.60 (n/a)134.08 (n/a)132.10 (n/a)120.90 (n/a)14.43 (n/a)

layer_norm_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-24.48%)0.06 (-15.22%)0.06 (-7.62%)0.05 (+2.87%)0.00 (-70.35%)154.40 (-2.77%)143.24 (+12.88%)143.30 (+8.31%)125.00 (+32.42%)10.84 (-60.68%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.07 (n/a)0.06 (n/a)0.05 (n/a)0.02 (n/a)158.80 (n/a)126.90 (n/a)132.30 (n/a)94.40 (n/a)27.56 (n/a)

layer_norm_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+4.40%)0.07 (+5.40%)0.07 (+0.90%)0.06 (+3.38%)0.01 (+26.26%)148.90 (-3.25%)122.60 (-4.28%)124.20 (-0.88%)97.90 (-4.21%)20.42 (+15.46%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)153.90 (n/a)128.08 (n/a)125.30 (n/a)102.20 (n/a)17.69 (n/a)

layer_norm_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-13.11%)0.06 (-16.10%)0.06 (-21.44%)0.05 (-5.54%)0.01 (-25.31%)161.90 (+5.82%)141.30 (+18.38%)142.90 (+27.36%)116.10 (+15.06%)15.46 (-13.01%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)153.00 (n/a)119.36 (n/a)112.20 (n/a)100.90 (n/a)17.77 (n/a)

layer_norm_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-14.63%)0.06 (-10.70%)0.06 (-8.19%)0.05 (-8.25%)0.01 (-19.91%)156.00 (+9.01%)135.78 (+11.77%)131.80 (+8.93%)119.90 (+17.20%)13.43 (+2.28%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)143.10 (n/a)121.48 (n/a)121.00 (n/a)102.30 (n/a)13.13 (n/a)

layer_norm_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.06 (+5.14%)0.05 (-4.42%)0.05 (-3.21%)0.05 (-9.28%)0.01 (+82.37%)175.60 (+10.23%)156.98 (+5.47%)153.40 (+3.30%)130.40 (-4.89%)16.31 (+91.54%)
130b6ea — 2025-12-05 21:15:490.06 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.00 (n/a)159.30 (n/a)148.84 (n/a)148.50 (n/a)137.10 (n/a)8.52 (n/a)

matrix_vector_mul_128x128_32_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:400.39 (+4.78%)0.32 (+6.94%)0.32 (+6.23%)0.22 (-1.83%)0.05 (-6.18%)
130b6ea — 2025-12-05 21:15:490.37 (n/a)0.30 (n/a)0.30 (n/a)0.23 (n/a)0.06 (n/a)

matrix_vector_mul_2048x8192_1_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:4013.28 (-2.12%)13.06 (-0.70%)13.13 (+0.28%)12.61 (+0.69%)0.24 (-36.72%)
130b6ea — 2025-12-05 21:15:4913.57 (n/a)13.15 (n/a)13.10 (n/a)12.52 (n/a)0.37 (n/a)

matrix_vector_mul_2048x8192_1_2col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:4024.84 (-0.89%)24.42 (+3.00%)24.60 (+1.95%)23.82 (+13.95%)0.40 (-73.03%)
130b6ea — 2025-12-05 21:15:4925.06 (n/a)23.71 (n/a)24.13 (n/a)20.90 (n/a)1.48 (n/a)

matrix_vector_mul_2048x8192_1_4col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:4041.09 (-2.28%)39.85 (-1.62%)39.41 (-2.66%)39.12 (+0.37%)0.76 (-21.70%)
130b6ea — 2025-12-05 21:15:4942.05 (n/a)40.50 (n/a)40.49 (n/a)38.98 (n/a)0.97 (n/a)

matrix_vector_mul_2048x8192_1_8col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:4045.26 (-2.93%)44.08 (-1.65%)43.80 (-2.19%)43.50 (+2.09%)0.61 (-61.54%)
130b6ea — 2025-12-05 21:15:4946.62 (n/a)44.82 (n/a)44.78 (n/a)42.60 (n/a)1.60 (n/a)

matrix_vector_mul_8192x2048_4_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:4013.31 (+0.18%)12.95 (-1.87%)13.01 (-1.33%)12.60 (-3.73%)0.27 (+260.37%)
130b6ea — 2025-12-05 21:15:4913.29 (n/a)13.19 (n/a)13.19 (n/a)13.09 (n/a)0.08 (n/a)

matrix_vector_mul_8192x2048_4_2col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:4024.00 (-6.16%)23.38 (-3.48%)23.93 (-0.40%)21.16 (-10.27%)1.11 (+59.58%)
130b6ea — 2025-12-05 21:15:4925.58 (n/a)24.23 (n/a)24.03 (n/a)23.58 (n/a)0.70 (n/a)

matrix_vector_mul_8192x2048_4_4col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:4040.79 (-0.38%)39.03 (-1.79%)38.85 (-2.31%)38.14 (-0.08%)0.94 (-9.44%)
130b6ea — 2025-12-05 21:15:4940.94 (n/a)39.74 (n/a)39.76 (n/a)38.17 (n/a)1.04 (n/a)

matrix_vector_mul_8192x2048_4_8col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
c6586c8 — 2025-12-05 23:36:4046.87 (+0.35%)45.77 (+1.92%)46.25 (+3.45%)43.70 (-0.24%)1.19 (+23.29%)
130b6ea — 2025-12-05 21:15:4946.71 (n/a)44.91 (n/a)44.71 (n/a)43.80 (n/a)0.96 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.05 (-20.09%)0.05 (-18.78%)0.05 (-20.49%)0.05 (-3.15%)0.00 (-65.77%)178.90 (+3.23%)167.60 (+21.50%)168.00 (+25.75%)154.40 (+25.12%)8.04 (-56.32%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)173.30 (n/a)137.94 (n/a)133.60 (n/a)123.40 (n/a)18.40 (n/a)

mem_copy_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (-16.99%)0.05 (-12.47%)0.05 (-7.32%)0.04 (-11.17%)0.00 (-29.61%)197.00 (+12.57%)171.00 (+13.70%)164.00 (+7.89%)153.00 (+20.47%)17.49 (-4.32%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)175.00 (n/a)150.40 (n/a)152.00 (n/a)127.00 (n/a)18.28 (n/a)

mem_copy_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (+5.56%)0.05 (-1.59%)0.05 (-1.39%)0.03 (-14.07%)0.01 (+25.03%)263.00 (+16.37%)166.00 (+4.53%)144.00 (+1.41%)126.00 (-5.26%)49.38 (+43.17%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)226.00 (n/a)158.80 (n/a)142.00 (n/a)133.00 (n/a)34.49 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-13.47%)0.06 (-17.45%)0.06 (-13.52%)0.04 (-18.23%)0.01 (-12.76%)183.70 (+22.30%)147.64 (+21.33%)145.20 (+15.61%)113.10 (+15.53%)23.04 (+24.15%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)150.20 (n/a)121.68 (n/a)125.60 (n/a)97.90 (n/a)18.56 (n/a)

mem_copy_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (-15.32%)0.05 (-17.09%)0.05 (-23.81%)0.04 (-15.54%)0.01 (-18.18%)193.00 (+18.40%)160.80 (+20.36%)168.00 (+31.25%)124.00 (+18.10%)22.92 (+11.39%)
0999a9a — 2025-11-13 19:24:080.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)163.00 (n/a)133.60 (n/a)128.00 (n/a)105.00 (n/a)20.58 (n/a)

mem_copy_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (-23.02%)0.05 (-6.98%)0.05 (-2.13%)0.04 (-15.31%)0.01 (-33.29%)196.00 (+18.07%)147.00 (+6.52%)141.00 (+2.17%)126.00 (+29.90%)25.27 (+8.13%)
0999a9a — 2025-11-13 19:24:080.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)166.00 (n/a)138.00 (n/a)138.00 (n/a)97.00 (n/a)23.37 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.06 (-42.78%)0.06 (-33.36%)0.06 (-30.18%)0.05 (-30.44%)0.00 (-74.88%)157.90 (+43.68%)145.92 (+48.44%)144.40 (+43.25%)138.70 (+74.69%)6.77 (-35.78%)
130b6ea — 2025-12-05 21:15:490.10 (n/a)0.08 (n/a)0.08 (n/a)0.07 (n/a)0.01 (n/a)109.90 (n/a)98.30 (n/a)100.80 (n/a)79.40 (n/a)10.54 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-27.86%)0.06 (-18.68%)0.06 (-20.25%)0.06 (+3.11%)0.01 (-61.92%)145.60 (-3.00%)137.12 (+19.65%)140.10 (+25.43%)116.30 (+38.62%)10.81 (-49.10%)
130b6ea — 2025-12-05 21:15:490.10 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)150.10 (n/a)114.60 (n/a)111.70 (n/a)83.90 (n/a)21.23 (n/a)

mem_copy_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.07 (+11.97%)0.05 (+3.09%)0.05 (+2.10%)0.03 (-11.35%)0.01 (+41.65%)229.00 (+12.81%)156.00 (-0.51%)143.00 (-2.05%)117.00 (-10.69%)38.21 (+48.71%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)203.00 (n/a)156.80 (n/a)146.00 (n/a)131.00 (n/a)25.69 (n/a)

mem_copy_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (+9.02%)0.05 (+5.59%)0.05 (+0.00%)0.04 (+3.63%)0.01 (+25.48%)193.00 (-3.50%)147.40 (-4.66%)147.00 (+0.00%)122.00 (-8.27%)25.24 (+7.48%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)200.00 (n/a)154.60 (n/a)147.00 (n/a)133.00 (n/a)23.48 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (-15.51%)0.06 (-14.07%)0.06 (-7.90%)0.04 (-30.04%)0.01 (+15.51%)185.40 (+42.95%)134.78 (+18.96%)126.90 (+8.55%)107.00 (+18.36%)28.81 (+91.57%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)129.70 (n/a)113.30 (n/a)116.90 (n/a)90.40 (n/a)15.04 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (+12.02%)0.07 (-2.45%)0.06 (-8.09%)0.06 (-1.09%)0.01 (+25.74%)136.00 (+1.12%)120.42 (+3.24%)127.10 (+8.82%)87.50 (-10.71%)17.31 (+10.18%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)134.50 (n/a)116.64 (n/a)116.80 (n/a)98.00 (n/a)15.71 (n/a)

mem_copy_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (-13.48%)0.05 (-17.79%)0.05 (-15.33%)0.03 (-25.09%)0.01 (+6.67%)267.00 (+33.50%)176.80 (+24.51%)150.00 (+18.11%)141.00 (+15.57%)47.41 (+61.41%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)200.00 (n/a)142.00 (n/a)127.00 (n/a)122.00 (n/a)29.37 (n/a)

mem_copy_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (+13.38%)0.05 (+23.85%)0.05 (+37.91%)0.05 (+30.72%)0.00 (-56.85%)166.00 (-23.50%)153.40 (-20.60%)153.00 (-27.49%)142.00 (-11.80%)7.63 (-70.60%)
0999a9a — 2025-11-13 19:24:080.05 (n/a)0.04 (n/a)0.04 (n/a)0.04 (n/a)0.01 (n/a)217.00 (n/a)193.20 (n/a)211.00 (n/a)161.00 (n/a)25.96 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (-12.96%)0.06 (-19.01%)0.06 (-20.66%)0.05 (-34.83%)0.01 (+66.93%)180.00 (+53.45%)136.36 (+26.49%)136.80 (+26.08%)108.10 (+14.88%)24.65 (+197.22%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.08 (n/a)0.08 (n/a)0.07 (n/a)0.01 (n/a)117.30 (n/a)107.80 (n/a)108.50 (n/a)94.10 (n/a)8.29 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (-15.62%)0.06 (-8.80%)0.06 (-11.13%)0.05 (-8.57%)0.01 (-18.26%)149.30 (+9.38%)132.24 (+9.42%)140.80 (+12.55%)108.50 (+18.45%)17.17 (+9.21%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)136.50 (n/a)120.86 (n/a)125.10 (n/a)91.60 (n/a)15.72 (n/a)

mha

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.18 (+0.03%)0.18 (-0.04%)0.18 (+0.01%)0.18 (-0.25%)0.00 (+167.44%)47373.90 (+0.25%)47230.52 (+0.04%)47205.50 (-0.01%)47152.90 (-0.03%)77.02 (+168.30%)
130b6ea — 2025-12-05 21:15:490.18 (n/a)0.18 (n/a)0.18 (n/a)0.18 (n/a)0.00 (n/a)47253.90 (n/a)47213.70 (n/a)47210.90 (n/a)47167.70 (n/a)28.71 (n/a)

relu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (+18.66%)0.07 (-0.96%)0.06 (-7.53%)0.06 (+3.19%)0.01 (+69.38%)132.10 (-3.08%)120.32 (+2.38%)127.80 (+8.21%)87.40 (-15.80%)16.76 (+38.34%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)136.30 (n/a)117.52 (n/a)118.10 (n/a)103.80 (n/a)12.12 (n/a)

relu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+11.79%)0.06 (+3.96%)0.06 (+1.03%)0.06 (+9.30%)0.01 (+19.90%)146.10 (-8.46%)131.82 (-3.68%)134.60 (-1.03%)109.00 (-10.51%)12.34 (-4.85%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)159.60 (n/a)136.86 (n/a)136.00 (n/a)121.80 (n/a)12.97 (n/a)

relu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+15.28%)0.07 (+16.08%)0.07 (+17.11%)0.06 (+22.48%)0.01 (+1.52%)140.50 (-18.36%)117.80 (-14.23%)112.60 (-14.63%)105.30 (-13.26%)12.91 (-29.27%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)172.10 (n/a)137.34 (n/a)131.90 (n/a)121.40 (n/a)18.25 (n/a)

relu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (+5.91%)0.06 (-1.75%)0.06 (-14.30%)0.05 (-3.43%)0.01 (+15.32%)171.60 (+3.56%)143.28 (+2.31%)147.10 (+16.65%)116.40 (-5.52%)21.53 (+11.28%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)165.70 (n/a)140.04 (n/a)126.10 (n/a)123.20 (n/a)19.35 (n/a)

rms_norm_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-1.10%)0.06 (-2.94%)0.06 (-3.66%)0.05 (-7.91%)0.01 (+16.85%)159.30 (+8.59%)139.06 (+3.48%)141.70 (+3.73%)114.10 (+1.06%)16.05 (+26.99%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.06 (n/a)0.01 (n/a)146.70 (n/a)134.38 (n/a)136.60 (n/a)112.90 (n/a)12.64 (n/a)

rms_norm_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (+19.73%)0.08 (+21.00%)0.08 (+19.08%)0.07 (+34.02%)0.01 (-10.18%)117.90 (-25.38%)103.34 (-18.20%)98.20 (-16.07%)89.30 (-16.46%)10.95 (-43.04%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)158.00 (n/a)126.34 (n/a)117.00 (n/a)106.90 (n/a)19.22 (n/a)

rms_norm_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (-2.06%)0.07 (+7.14%)0.06 (-0.47%)0.06 (+18.07%)0.01 (-24.37%)146.00 (-15.31%)126.88 (-8.88%)135.80 (+0.44%)95.90 (+2.13%)18.44 (-35.73%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)172.40 (n/a)139.24 (n/a)135.20 (n/a)93.90 (n/a)28.69 (n/a)

rms_norm_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-0.56%)0.07 (+16.08%)0.07 (+20.42%)0.06 (+27.95%)0.00 (-55.27%)132.50 (-21.83%)117.18 (-15.76%)115.60 (-17.01%)110.40 (+0.55%)7.97 (-64.46%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)169.50 (n/a)139.10 (n/a)139.30 (n/a)109.80 (n/a)22.44 (n/a)

rms_norm_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (+25.15%)0.06 (-0.99%)0.05 (-15.49%)0.05 (-21.46%)0.02 (+261.82%)173.80 (+27.33%)134.86 (+7.03%)152.20 (+18.35%)89.00 (-20.11%)31.71 (+267.29%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.07 (n/a)0.06 (n/a)0.06 (n/a)0.00 (n/a)136.50 (n/a)126.00 (n/a)128.60 (n/a)111.40 (n/a)8.63 (n/a)

rms_norm_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (+3.04%)0.07 (+3.14%)0.07 (-1.96%)0.06 (+15.21%)0.00 (-38.01%)126.70 (-13.16%)120.46 (-3.51%)121.80 (+2.01%)113.50 (-2.99%)5.66 (-47.90%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)145.90 (n/a)124.84 (n/a)119.40 (n/a)117.00 (n/a)10.86 (n/a)

rms_norm_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (+7.14%)0.06 (+11.90%)0.06 (+9.93%)0.04 (+7.36%)0.01 (-2.79%)187.20 (-6.87%)140.60 (-11.19%)135.00 (-9.03%)115.70 (-6.69%)25.36 (-15.11%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.05 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)201.00 (n/a)158.32 (n/a)148.40 (n/a)124.00 (n/a)29.87 (n/a)

rms_norm_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (+29.12%)0.07 (+24.00%)0.07 (+20.06%)0.06 (+27.69%)0.01 (+19.89%)139.00 (-21.65%)117.84 (-19.54%)117.70 (-16.70%)95.40 (-22.56%)13.96 (-28.34%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)177.40 (n/a)146.46 (n/a)141.30 (n/a)123.20 (n/a)19.48 (n/a)

rope_1_cols_2_channels_4096_tile_4096_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.21 (-22.58%)0.17 (-19.25%)0.16 (-16.72%)0.15 (-13.44%)0.02 (-40.78%)163.50 (+15.47%)146.76 (+22.12%)157.20 (+20.09%)119.30 (+29.11%)17.65 (-10.54%)
130b6ea — 2025-12-05 21:15:490.27 (n/a)0.21 (n/a)0.19 (n/a)0.17 (n/a)0.04 (n/a)141.60 (n/a)120.18 (n/a)130.90 (n/a)92.40 (n/a)19.73 (n/a)

rope_2_cols_2_channels_4096_tile_2048_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.20 (-25.30%)0.17 (-10.50%)0.18 (-11.85%)0.15 (+2.14%)0.02 (-60.94%)167.90 (-2.10%)141.78 (+7.88%)139.70 (+13.48%)125.50 (+33.80%)14.14 (-48.33%)
130b6ea — 2025-12-05 21:15:490.26 (n/a)0.20 (n/a)0.20 (n/a)0.14 (n/a)0.04 (n/a)171.50 (n/a)131.42 (n/a)123.10 (n/a)93.80 (n/a)27.37 (n/a)

rope_4_cols_2_channels_4096_tile_1024_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.25 (-7.74%)0.18 (-14.90%)0.15 (-25.16%)0.15 (-14.15%)0.04 (+12.30%)167.00 (+16.54%)144.02 (+19.24%)159.40 (+33.61%)98.10 (+8.40%)25.41 (+43.56%)
130b6ea — 2025-12-05 21:15:490.27 (n/a)0.21 (n/a)0.21 (n/a)0.17 (n/a)0.03 (n/a)143.30 (n/a)120.78 (n/a)119.30 (n/a)90.50 (n/a)17.70 (n/a)

rope_8_cols_2_channels_4096_tile_512_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.20 (-11.38%)0.16 (-10.71%)0.15 (-13.01%)0.12 (-18.51%)0.03 (+16.92%)208.90 (+22.67%)163.28 (+14.44%)164.20 (+14.91%)121.30 (+12.84%)34.70 (+61.49%)
130b6ea — 2025-12-05 21:15:490.23 (n/a)0.18 (n/a)0.17 (n/a)0.14 (n/a)0.03 (n/a)170.30 (n/a)142.68 (n/a)142.90 (n/a)107.50 (n/a)21.49 (n/a)

sigmoid_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.10 (+13.61%)0.08 (+8.32%)0.07 (+12.15%)0.06 (+3.01%)0.01 (+28.02%)130.80 (-2.97%)111.62 (-7.08%)115.70 (-10.86%)82.00 (-12.02%)16.53 (+6.10%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.07 (n/a)0.06 (n/a)0.06 (n/a)0.01 (n/a)134.80 (n/a)120.12 (n/a)129.80 (n/a)93.20 (n/a)15.58 (n/a)

sigmoid_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.10 (+22.37%)0.07 (+4.93%)0.07 (+5.62%)0.05 (-8.25%)0.02 (+81.77%)149.90 (+8.94%)119.88 (-1.69%)123.30 (-5.30%)79.50 (-18.29%)24.23 (+57.65%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.06 (n/a)0.06 (n/a)0.01 (n/a)137.60 (n/a)121.94 (n/a)130.20 (n/a)97.30 (n/a)15.37 (n/a)

sigmoid_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (+8.94%)0.07 (+11.57%)0.07 (+17.85%)0.06 (+13.83%)0.01 (+20.37%)137.70 (-12.13%)114.68 (-9.87%)110.60 (-15.18%)89.90 (-8.17%)19.81 (+1.26%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)156.70 (n/a)127.24 (n/a)130.40 (n/a)97.90 (n/a)19.56 (n/a)

sigmoid_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+18.88%)0.07 (+17.53%)0.06 (+8.49%)0.05 (+9.12%)0.01 (+37.18%)173.50 (-8.39%)128.52 (-13.84%)133.00 (-7.83%)97.70 (-15.92%)27.97 (+1.65%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)189.40 (n/a)149.16 (n/a)144.30 (n/a)116.20 (n/a)27.52 (n/a)

silu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-14.17%)0.06 (-13.47%)0.06 (-17.65%)0.06 (+3.09%)0.01 (-43.57%)139.90 (-2.98%)132.60 (+14.33%)136.80 (+21.49%)113.60 (+16.51%)9.67 (-38.23%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)144.20 (n/a)115.98 (n/a)112.60 (n/a)97.50 (n/a)15.66 (n/a)

silu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (-22.35%)0.07 (-15.81%)0.07 (-26.23%)0.06 (+38.15%)0.01 (-64.46%)143.10 (-27.58%)122.20 (+7.01%)124.10 (+35.48%)100.00 (+28.87%)14.75 (-67.01%)
130b6ea — 2025-12-05 21:15:490.11 (n/a)0.08 (n/a)0.09 (n/a)0.04 (n/a)0.02 (n/a)197.60 (n/a)114.20 (n/a)91.60 (n/a)77.60 (n/a)44.72 (n/a)

silu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+10.84%)0.07 (+2.06%)0.06 (-9.87%)0.06 (+30.76%)0.01 (-13.02%)144.60 (-23.49%)126.86 (-3.89%)136.40 (+10.98%)97.90 (-9.77%)17.71 (-40.37%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.06 (n/a)0.07 (n/a)0.04 (n/a)0.01 (n/a)189.00 (n/a)132.00 (n/a)122.90 (n/a)108.50 (n/a)29.70 (n/a)

silu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.07 (-8.50%)0.06 (-7.67%)0.06 (-4.90%)0.05 (-1.91%)0.01 (-35.38%)160.80 (+1.97%)140.02 (+6.45%)140.50 (+5.09%)112.00 (+9.27%)15.91 (-30.34%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)157.70 (n/a)131.54 (n/a)133.70 (n/a)102.50 (n/a)22.84 (n/a)

softmax_1_cols_2_channels_4096_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.14 (-20.42%)0.12 (-21.17%)0.12 (-22.11%)0.10 (-20.36%)0.01 (-38.92%)164.20 (+25.63%)142.02 (+25.97%)139.40 (+28.36%)121.10 (+25.62%)14.08 (-4.64%)
130b6ea — 2025-12-05 21:15:490.17 (n/a)0.15 (n/a)0.15 (n/a)0.13 (n/a)0.02 (n/a)130.70 (n/a)112.74 (n/a)108.60 (n/a)96.40 (n/a)14.77 (n/a)

softmax_2_cols_2_channels_4096_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.18 (+2.92%)0.14 (-8.23%)0.14 (-13.61%)0.13 (-0.44%)0.02 (+20.08%)128.80 (+0.47%)115.66 (+9.38%)117.70 (+15.73%)92.10 (-2.85%)13.21 (+13.55%)
130b6ea — 2025-12-05 21:15:490.17 (n/a)0.16 (n/a)0.16 (n/a)0.13 (n/a)0.02 (n/a)128.20 (n/a)105.74 (n/a)101.70 (n/a)94.80 (n/a)11.63 (n/a)

softmax_2_cols_2_channels_4096_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.18 (+22.15%)0.13 (-7.56%)0.10 (-26.98%)0.09 (-28.51%)0.04 (+691.03%)175.50 (+39.84%)141.80 (+17.15%)167.20 (+36.94%)92.90 (-18.15%)37.32 (+823.05%)
130b6ea — 2025-12-05 21:15:490.14 (n/a)0.14 (n/a)0.13 (n/a)0.13 (n/a)0.00 (n/a)125.50 (n/a)121.04 (n/a)122.10 (n/a)113.50 (n/a)4.04 (n/a)

swiglu

No metrics available.

swiglu_decode_1x2048x2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.00 (+17.86%)0.00 (+18.22%)0.00 (+37.73%)0.00 (-2.29%)0.00 (+152.74%)5308.90 (+2.35%)4128.28 (-12.55%)3449.50 (-27.39%)3406.70 (-15.15%)859.54 (+120.26%)
130b6ea — 2025-12-05 21:15:490.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)5187.20 (n/a)4720.84 (n/a)4750.90 (n/a)4015.10 (n/a)390.23 (n/a)

tanh_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (-15.43%)0.08 (+6.16%)0.08 (+16.08%)0.07 (+18.90%)0.00 (-77.77%)115.50 (-15.88%)107.90 (-8.12%)106.30 (-13.79%)104.70 (+18.31%)4.01 (-77.91%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)137.30 (n/a)117.44 (n/a)123.30 (n/a)88.50 (n/a)18.13 (n/a)

tanh_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+7.56%)0.06 (-1.82%)0.05 (-6.74%)0.05 (+1.73%)0.01 (+25.88%)162.50 (-1.75%)140.40 (+2.59%)150.70 (+7.26%)104.30 (-7.04%)20.32 (+12.93%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)165.40 (n/a)136.86 (n/a)140.50 (n/a)112.20 (n/a)17.99 (n/a)

tanh_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.09 (-1.76%)0.06 (-6.16%)0.06 (-13.97%)0.05 (-7.16%)0.01 (-2.51%)166.80 (+7.68%)132.70 (+6.50%)138.00 (+16.26%)93.70 (+1.74%)24.83 (+1.71%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)154.90 (n/a)124.60 (n/a)118.70 (n/a)92.10 (n/a)24.41 (n/a)

tanh_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.06 (-31.19%)0.06 (-22.77%)0.06 (-21.23%)0.05 (-17.72%)0.01 (-58.73%)167.10 (+21.53%)146.86 (+26.98%)147.50 (+26.94%)128.50 (+45.36%)13.43 (-28.63%)
130b6ea — 2025-12-05 21:15:490.09 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)137.50 (n/a)115.66 (n/a)116.20 (n/a)88.40 (n/a)18.82 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:403.96 (+0.96%)3.56 (-3.20%)3.68 (+0.75%)3.19 (-7.73%)0.29 (+90.47%)164.50 (+8.37%)148.14 (+3.81%)142.30 (-0.77%)132.30 (-0.90%)12.19 (+106.78%)
130b6ea — 2025-12-05 21:15:493.93 (n/a)3.68 (n/a)3.66 (n/a)3.45 (n/a)0.15 (n/a)151.80 (n/a)142.70 (n/a)143.40 (n/a)133.50 (n/a)5.90 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:404.97 (-2.78%)4.05 (+5.13%)3.96 (+16.47%)3.43 (+4.47%)0.50 (-29.68%)152.80 (-4.26%)131.24 (-6.38%)132.40 (-14.14%)105.60 (+2.92%)15.04 (-33.75%)
130b6ea — 2025-12-05 21:15:495.11 (n/a)3.85 (n/a)3.40 (n/a)3.29 (n/a)0.71 (n/a)159.60 (n/a)140.18 (n/a)154.20 (n/a)102.60 (n/a)22.70 (n/a)

weighted_rms_norm_1_cols_2_channels_2048_weights_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+6.24%)0.07 (+11.64%)0.07 (+9.71%)0.06 (+34.65%)0.01 (-33.00%)135.70 (-25.73%)120.74 (-12.44%)119.90 (-8.82%)100.30 (-5.91%)11.87 (-54.36%)
130b6ea — 2025-12-05 21:15:490.08 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)182.70 (n/a)137.90 (n/a)131.50 (n/a)106.60 (n/a)26.01 (n/a)

weighted_rms_norm_2_cols_2_channels_2048_weights_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+12.55%)0.07 (+0.57%)0.06 (-9.03%)0.05 (-12.17%)0.01 (+65.36%)169.80 (+13.88%)128.70 (+1.59%)131.60 (+9.85%)100.20 (-11.17%)24.37 (+67.88%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)149.10 (n/a)126.68 (n/a)119.80 (n/a)112.80 (n/a)14.52 (n/a)

weighted_rms_norm_4_cols_2_channels_2048_weights_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+9.61%)0.06 (+5.07%)0.06 (+3.78%)0.05 (+8.76%)0.01 (+2.75%)160.50 (-8.08%)130.02 (-5.08%)130.30 (-3.62%)103.00 (-8.77%)20.56 (-11.67%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)174.60 (n/a)136.98 (n/a)135.20 (n/a)112.90 (n/a)23.27 (n/a)

weighted_rms_norm_8_cols_2_channels_2048_weights_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
c6586c8 — 2025-12-05 23:36:400.08 (+14.83%)0.06 (+7.52%)0.05 (+0.86%)0.05 (+5.38%)0.01 (+47.26%)171.80 (-5.08%)149.18 (-5.70%)159.60 (-0.87%)107.70 (-12.93%)24.04 (+25.44%)
130b6ea — 2025-12-05 21:15:490.07 (n/a)0.05 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)181.00 (n/a)158.20 (n/a)161.00 (n/a)123.70 (n/a)19.16 (n/a)

@hunhoffe hunhoffe marked this pull request as ready for review December 9, 2025 17:07
@github-actions
Copy link
Contributor

📊 Test Results for Test Example Applications

68acccf (2025_12_16_23_03_25)

IRONCLAD

Tested on 2025_12_16_23_03_25 at commit 68acccf.

Test Checks TTFT (mean)TPS (mean)
llama_3.2_1b_prompt_13_tokens_1 ✅ 1/1 0.64 0.00
llama_3.2_1b_prompt_13_tokens_40 ✅ 1/1 0.62 5.02
llama_3.2_1b_prompt_2048_tokens_1 ✅ 1/1 4.38 0.00
llama_3.2_1b_prompt_2048_tokens_40 ✅ 1/1 4.37 4.64
📈 Trends (vs main branch) for Test Example Applications

68acccf (2025_12_16_23_03_25)

IRONCLAD Trends

llama_3.2_1b

Commit/Date Num Tokens (max)Num Tokens (mean)Num Tokens (median)Num Tokens (min)Num Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)Total (max)Total (mean)Total (median)Total (min)Total (stddev)
130b6ea — 2025-12-05 21:33:1240.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)0.00 (n/a)4.71 (-0.42%)4.64 (-0.09%)4.64 (+0.65%)4.55 (-0.22%)0.05 (-17.66%)4.41 (-0.34%)4.39 (-0.19%)4.38 (-0.33%)4.37 (-0.15%)0.01 (-25.90%)12.96 (-0.00%)12.80 (+0.07%)12.80 (-0.23%)12.67 (+0.44%)0.09 (-21.12%)
0a6c11c — 2025-12-03 23:35:1540.00 (n/a)40.00 (n/a)40.00 (n/a)40.00 (n/a)0.00 (n/a)4.73 (n/a)4.64 (n/a)4.61 (n/a)4.56 (n/a)0.06 (n/a)4.42 (n/a)4.40 (n/a)4.40 (n/a)4.37 (n/a)0.02 (n/a)12.96 (n/a)12.79 (n/a)12.83 (n/a)12.62 (n/a)0.12 (n/a)

llama_3.2_1b_prompt_13_tokens_1

Commit/Date Num_Tokens (max)Num_Tokens (mean)Num_Tokens (median)Num_Tokens (min)Num_Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)
68acccf — 2025-12-16 22:58:191.00 (+0.00%)1.00 (+0.00%)1.00 (+0.00%)1.00 (+0.00%)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.64 (+0.36%)0.64 (+0.36%)0.64 (+0.36%)0.64 (+0.36%)0.00 (n/a)
87e12ba — 2025-12-16 22:04:541.00 (n/a)1.00 (n/a)1.00 (n/a)1.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.63 (n/a)0.63 (n/a)0.63 (n/a)0.63 (n/a)0.00 (n/a)

llama_3.2_1b_prompt_13_tokens_40

Commit/Date Num_Tokens (max)Num_Tokens (mean)Num_Tokens (median)Num_Tokens (min)Num_Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)
68acccf — 2025-12-16 22:58:1940.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)0.00 (n/a)5.02 (+2.66%)5.02 (+2.66%)5.02 (+2.66%)5.02 (+2.66%)0.00 (n/a)0.62 (-2.25%)0.62 (-2.25%)0.62 (-2.25%)0.62 (-2.25%)0.00 (n/a)
87e12ba — 2025-12-16 22:04:5440.00 (n/a)40.00 (n/a)40.00 (n/a)40.00 (n/a)0.00 (n/a)4.89 (n/a)4.89 (n/a)4.89 (n/a)4.89 (n/a)0.00 (n/a)0.64 (n/a)0.64 (n/a)0.64 (n/a)0.64 (n/a)0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_1

Commit/Date Num_Tokens (max)Num_Tokens (mean)Num_Tokens (median)Num_Tokens (min)Num_Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)
68acccf — 2025-12-16 22:58:191.00 (+0.00%)1.00 (+0.00%)1.00 (+0.00%)1.00 (+0.00%)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)4.38 (-0.29%)4.38 (-0.29%)4.38 (-0.29%)4.38 (-0.29%)0.00 (n/a)
87e12ba — 2025-12-16 22:04:541.00 (n/a)1.00 (n/a)1.00 (n/a)1.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)4.39 (n/a)4.39 (n/a)4.39 (n/a)4.39 (n/a)0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_40

Commit/Date Num_Tokens (max)Num_Tokens (mean)Num_Tokens (median)Num_Tokens (min)Num_Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)
68acccf — 2025-12-16 22:58:1940.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)0.00 (n/a)4.64 (+2.65%)4.64 (+2.65%)4.64 (+2.65%)4.64 (+2.65%)0.00 (n/a)4.37 (+0.24%)4.37 (+0.24%)4.37 (+0.24%)4.37 (+0.24%)0.00 (n/a)
87e12ba — 2025-12-16 22:04:5440.00 (n/a)40.00 (n/a)40.00 (n/a)40.00 (n/a)0.00 (n/a)4.52 (n/a)4.52 (n/a)4.52 (n/a)4.52 (n/a)0.00 (n/a)4.36 (n/a)4.36 (n/a)4.36 (n/a)4.36 (n/a)0.00 (n/a)

@github-actions
Copy link
Contributor

📊 Test Results for Small Benchmark/Test Suite

68acccf (2025_12_16_23_12_51)

IRONCLAD

Tested on 2025_12_16_23_12_51 at commit 68acccf.

Test Checks Latency (mean)Bandwidth (mean)GFLOP/s (mean)
axpy_1_cols_2_channels_2048_tile_2048_3.0_0 ✅ 5/5 133.36 0.09 n/a
axpy_2_cols_2_channels_2048_tile_1024_3.0_0 ✅ 5/5 152.76 0.08 n/a
axpy_4_cols_2_channels_2048_tile_512_3.0_0 ✅ 5/5 143.40 0.09 n/a
axpy_8_cols_2_channels_2048_tile_256_3.0_0 ✅ 5/5 164.70 0.08 n/a
dequant_1_cols_1_channels_2048_tile_2048_0 ✅ 5/5 133.36 0.04 n/a
dequant_1_cols_2_channels_2048_tile_1024_0 ✅ 5/5 137.44 0.04 n/a
dequant_2_cols_1_channels_2048_tile_1024_0 ✅ 5/5 135.68 0.04 n/a
dequant_2_cols_2_channels_2048_tile_512_0 ✅ 5/5 153.56 0.04 n/a
dequant_4_cols_1_channels_2048_tile_512_0 ✅ 5/5 130.96 0.04 n/a
dequant_4_cols_2_channels_2048_tile_256_0 ✅ 5/5 149.82 0.04 n/a
dequant_8_cols_1_channels_2048_tile_256_0 ✅ 5/5 185.26 0.03 n/a
dequant_8_cols_2_channels_2048_tile_128_0 ✅ 5/5 165.32 0.03 n/a
eltwise_add_1_cols_2_channels_2048_tile_2048 ✅ 5/5 137.36 0.09 n/a
eltwise_add_2_cols_2_channels_2048_tile_1024 ✅ 5/5 138.76 0.09 n/a
eltwise_add_4_cols_2_channels_2048_tile_512 ✅ 5/5 122.04 0.10 n/a
eltwise_add_8_cols_2_channels_2048_tile_256 ✅ 5/5 141.64 0.09 n/a
eltwise_mul_1_cols_2_channels_2048_tile_2048 ✅ 5/5 135.40 0.09 n/a
eltwise_mul_2_cols_2_channels_2048_tile_1024 ✅ 5/5 132.16 0.09 n/a
eltwise_mul_4_cols_2_channels_2048_tile_512 ✅ 5/5 145.06 0.09 n/a
eltwise_mul_8_cols_2_channels_2048_tile_256 ✅ 5/5 155.80 0.08 n/a
gelu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 139.32 0.06 n/a
gelu_1_cols_2_channels_2048_tile_1024 ✅ 5/5 142.18 0.07 n/a
gelu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 133.88 0.06 n/a
gelu_2_cols_2_channels_2048_tile_512 ✅ 5/5 148.52 0.06 n/a
gelu_4_cols_1_channels_2048_tile_512 ✅ 5/5 150.04 0.06 n/a
gelu_4_cols_2_channels_2048_tile_256 ✅ 5/5 161.54 0.05 n/a
gelu_8_cols_1_channels_2048_tile_256 ✅ 5/5 142.26 0.06 n/a
gelu_8_cols_2_channels_2048_tile_128 ✅ 5/5 169.76 0.05 n/a
gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0 ✅ 5/5 24560.44 1.02 n/a
gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0 ✅ 5/5 25943.22 0.97 n/a
gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0 ✅ 5/5 27586.74 0.91 n/a
layer_norm_1_cols_1_channels_2048_tile_2048 ✅ 5/5 141.66 0.06 n/a
layer_norm_1_cols_2_channels_2048_tile_1024 ✅ 5/5 146.22 0.06 n/a
layer_norm_2_cols_1_channels_2048_tile_1024 ✅ 5/5 138.52 0.06 n/a
layer_norm_2_cols_2_channels_2048_tile_512 ✅ 5/5 142.22 0.06 n/a
layer_norm_4_cols_1_channels_2048_tile_512 ✅ 5/5 156.30 0.05 n/a
layer_norm_4_cols_2_channels_2048_tile_256 ✅ 5/5 158.22 0.05 n/a
layer_norm_8_cols_1_channels_2048_tile_256 ✅ 5/5 157.66 0.05 n/a
layer_norm_8_cols_2_channels_2048_tile_128 ✅ 5/5 154.66 0.05 n/a
matrix_vector_mul_128x128_32_1col0 ✅ 5/5 n/a 0.24 n/a
matrix_vector_mul_2048x8192_1_1col0 ✅ 5/5 n/a 13.06 n/a
matrix_vector_mul_2048x8192_1_2col0 ✅ 5/5 n/a 24.28 n/a
matrix_vector_mul_2048x8192_1_4col0 ✅ 5/5 n/a 39.02 n/a
matrix_vector_mul_2048x8192_1_8col0 ✅ 5/5 n/a 40.99 n/a
matrix_vector_mul_8192x2048_4_1col0 ✅ 5/5 n/a 12.94 n/a
matrix_vector_mul_8192x2048_4_2col0 ✅ 5/5 n/a 22.66 n/a
matrix_vector_mul_8192x2048_4_4col0 ✅ 5/5 n/a 39.21 n/a
matrix_vector_mul_8192x2048_4_8col0 ✅ 5/5 n/a 40.43 n/a
mem_copy_16_cores_2_chans_2048_tile_128_False0 ✅ 5/5 157.32 0.05 n/a
mem_copy_1_cores_1_chans_2048_tile_2048_False0 ✅ 5/5 139.12 0.06 n/a
mem_copy_2_cores_1_chans_2048_tile_1024_False0 ✅ 5/5 141.30 0.06 n/a
mem_copy_2_cores_2_chans_2048_tile_1024_False0 ✅ 5/5 122.38 0.07 n/a
mem_copy_4_cores_1_chans_2048_tile_512_False0 ✅ 5/5 137.34 0.06 n/a
mem_copy_4_cores_2_chans_2048_tile_512_False0 ✅ 5/5 122.52 0.07 n/a
mem_copy_8_cores_1_chans_2048_tile_256_False0 ✅ 5/5 128.18 0.07 n/a
mem_copy_8_cores_2_chans_2048_tile_256_False0 ✅ 5/5 139.88 0.06 n/a
mha0 ✅ 5/5 47191.42 0.18 n/a
relu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 144.06 0.06 n/a
relu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 149.62 0.06 n/a
relu_4_cols_1_channels_2048_tile_512 ✅ 5/5 152.06 0.05 n/a
relu_8_cols_1_channels_2048_tile_256 ✅ 5/5 145.86 0.06 n/a
rms_norm_1_cols_1_channels_2048_tile_2048 ✅ 5/5 134.52 0.06 n/a
rms_norm_1_cols_2_channels_2048_tile_1024 ✅ 5/5 156.46 0.05 n/a
rms_norm_2_cols_1_channels_2048_tile_1024 ✅ 5/5 137.86 0.06 n/a
rms_norm_2_cols_2_channels_2048_tile_512 ✅ 5/5 148.98 0.06 n/a
rms_norm_4_cols_1_channels_2048_tile_512 ✅ 5/5 138.78 0.06 n/a
rms_norm_4_cols_2_channels_2048_tile_256 ✅ 5/5 137.56 0.06 n/a
rms_norm_8_cols_1_channels_2048_tile_256 ✅ 5/5 149.78 0.06 n/a
rms_norm_8_cols_2_channels_2048_tile_128 ✅ 5/5 145.74 0.06 n/a
rope_1_cols_2_channels_4096_tile_4096_0 ✅ 5/5 135.62 0.18 n/a
rope_2_cols_2_channels_4096_tile_2048_0 ✅ 5/5 139.78 0.18 n/a
rope_4_cols_2_channels_4096_tile_1024_0 ✅ 5/5 140.60 0.18 n/a
rope_8_cols_2_channels_4096_tile_512_0 ✅ 5/5 183.96 0.14 n/a
sigmoid_1_cols_1_channels_2048_tile_2048 ✅ 5/5 138.32 0.06 n/a
sigmoid_2_cols_1_channels_2048_tile_1024 ✅ 5/5 161.36 0.05 n/a
sigmoid_4_cols_1_channels_2048_tile_512 ✅ 5/5 156.20 0.05 n/a
sigmoid_8_cols_1_channels_2048_tile_256 ✅ 5/5 164.30 0.05 n/a
silu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 127.86 0.07 n/a
silu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 158.00 0.05 n/a
silu_4_cols_1_channels_2048_tile_512 ✅ 5/5 142.84 0.06 n/a
silu_8_cols_1_channels_2048_tile_256 ✅ 5/5 130.56 0.06 n/a
softmax_1_cols_2_channels_4096_tile_2048 ✅ 5/5 131.84 0.12 n/a
softmax_2_cols_2_channels_4096_tile_1024 ✅ 5/5 152.58 0.11 n/a
softmax_2_cols_2_channels_4096_tile_512 ✅ 5/5 136.06 0.12 n/a
swiglu_decode_1x2048x2048_0 ✅ 5/5 4205.48 0.00 n/a
tanh_1_cols_1_channels_2048_tile_2048 ✅ 5/5 146.90 0.06 n/a
tanh_2_cols_1_channels_2048_tile_1024 ✅ 5/5 150.82 0.05 n/a
tanh_4_cols_1_channels_2048_tile_512 ✅ 5/5 174.32 0.05 n/a
tanh_8_cols_1_channels_2048_tile_256 ✅ 5/5 161.32 0.05 n/a
transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0 ✅ 5/5 138.32 3.80 n/a
transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0 ✅ 5/5 147.38 3.65 n/a
weighted_rms_norm_1_cols_2_channels_2048_weights_2048 ✅ 5/5 133.08 0.06 n/a
weighted_rms_norm_2_cols_2_channels_2048_weights_1024 ✅ 5/5 148.82 0.06 n/a
weighted_rms_norm_4_cols_2_channels_2048_weights_512 ✅ 5/5 134.96 0.06 n/a
weighted_rms_norm_8_cols_2_channels_2048_weights_256 ✅ 5/5 170.32 0.05 n/a
📈 Trends (vs main branch) for Small Benchmark/Test Suite

68acccf (2025_12_16_23_12_51)

IRONCLAD Trends

axpy_1_cols_2_channels_2048_tile_2048_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.13 (-1.96%)0.11 (+2.52%)0.10 (+1.86%)0.08 (-1.91%)0.02 (-1.72%)162.40 (+1.95%)120.96 (-2.36%)117.00 (-1.85%)94.40 (+2.05%)23.11 (+4.09%)
0a6c11c — 2025-12-04 00:39:100.13 (n/a)0.10 (n/a)0.10 (n/a)0.08 (n/a)0.02 (n/a)159.30 (n/a)123.88 (n/a)119.20 (n/a)92.50 (n/a)22.21 (n/a)

axpy_1_cols_2_channels_2048_tile_2048_3.0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.11 (-22.20%)0.09 (-14.05%)0.10 (-11.31%)0.08 (-9.50%)0.01 (-41.65%)149.70 (+10.48%)133.36 (+15.10%)127.20 (+12.77%)114.60 (+28.48%)15.46 (-14.63%)
87e12ba — 2025-12-16 21:55:160.14 (n/a)0.11 (n/a)0.11 (n/a)0.09 (n/a)0.02 (n/a)135.50 (n/a)115.86 (n/a)112.80 (n/a)89.20 (n/a)18.11 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.11 (-26.77%)0.09 (-9.09%)0.09 (-10.81%)0.08 (+8.25%)0.01 (-55.42%)152.00 (-7.60%)135.46 (+6.28%)142.80 (+12.18%)115.00 (+36.58%)15.03 (-41.36%)
0a6c11c — 2025-12-04 00:39:100.15 (n/a)0.10 (n/a)0.10 (n/a)0.07 (n/a)0.02 (n/a)164.50 (n/a)127.46 (n/a)127.30 (n/a)84.20 (n/a)25.64 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.09 (-22.29%)0.08 (-17.40%)0.08 (-22.59%)0.07 (-4.96%)0.01 (-63.68%)172.10 (+5.20%)152.76 (+18.42%)152.90 (+29.25%)138.90 (+28.73%)12.37 (-49.84%)
87e12ba — 2025-12-16 21:55:160.11 (n/a)0.10 (n/a)0.10 (n/a)0.08 (n/a)0.02 (n/a)163.60 (n/a)129.00 (n/a)118.30 (n/a)107.90 (n/a)24.65 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.12 (-10.21%)0.10 (-8.41%)0.10 (-6.20%)0.07 (+3.52%)0.02 (-23.48%)166.70 (-3.36%)128.78 (+7.33%)121.80 (+6.65%)103.90 (+11.36%)22.95 (-18.82%)
0a6c11c — 2025-12-04 00:39:100.13 (n/a)0.11 (n/a)0.11 (n/a)0.07 (n/a)0.02 (n/a)172.50 (n/a)119.98 (n/a)114.20 (n/a)93.30 (n/a)28.27 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.10 (-21.10%)0.09 (-8.20%)0.08 (-6.40%)0.08 (+8.46%)0.01 (-61.31%)156.60 (-7.77%)143.40 (+6.35%)148.30 (+6.84%)127.30 (+26.79%)11.53 (-54.45%)
87e12ba — 2025-12-16 21:55:160.12 (n/a)0.09 (n/a)0.09 (n/a)0.07 (n/a)0.02 (n/a)169.80 (n/a)134.84 (n/a)138.80 (n/a)100.40 (n/a)25.31 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.10 (-16.19%)0.08 (-10.40%)0.09 (-2.09%)0.07 (-17.38%)0.01 (-10.61%)184.40 (+21.00%)153.18 (+11.99%)144.10 (+2.20%)127.10 (+19.34%)21.26 (+34.76%)
0a6c11c — 2025-12-04 00:39:100.12 (n/a)0.09 (n/a)0.09 (n/a)0.08 (n/a)0.01 (n/a)152.40 (n/a)136.78 (n/a)141.00 (n/a)106.50 (n/a)15.78 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.09 (-27.82%)0.08 (-30.49%)0.07 (-35.48%)0.07 (-29.80%)0.01 (-19.61%)187.90 (+42.46%)164.70 (+44.22%)170.60 (+54.95%)135.10 (+38.56%)20.42 (+56.35%)
87e12ba — 2025-12-16 21:55:160.13 (n/a)0.11 (n/a)0.11 (n/a)0.09 (n/a)0.01 (n/a)131.90 (n/a)114.20 (n/a)110.10 (n/a)97.50 (n/a)13.06 (n/a)

dequant_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.05 (-18.83%)0.04 (-5.21%)0.04 (-2.69%)0.03 (+37.10%)0.00 (-53.98%)160.40 (-27.06%)141.12 (-0.91%)136.40 (+2.79%)114.30 (+23.17%)17.37 (-59.06%)
0a6c11c — 2025-12-04 00:39:100.06 (n/a)0.04 (n/a)0.04 (n/a)0.02 (n/a)0.01 (n/a)219.90 (n/a)142.42 (n/a)132.70 (n/a)92.80 (n/a)42.43 (n/a)

dequant_1_cols_1_channels_2048_tile_2048_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.05 (-4.25%)0.04 (-5.94%)0.04 (-11.10%)0.03 (-14.07%)0.01 (+21.62%)169.40 (+16.35%)133.36 (+7.50%)133.00 (+12.43%)107.60 (+4.36%)24.87 (+44.47%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.04 (n/a)0.04 (n/a)0.04 (n/a)0.01 (n/a)145.60 (n/a)124.06 (n/a)118.30 (n/a)103.10 (n/a)17.21 (n/a)

dequant_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.06 (-13.28%)0.05 (-8.93%)0.06 (-13.59%)0.04 (+4.63%)0.01 (-39.10%)133.60 (-4.43%)106.08 (+5.85%)95.10 (+15.69%)87.90 (+15.35%)17.90 (-31.97%)
0a6c11c — 2025-12-04 00:39:100.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)139.80 (n/a)100.22 (n/a)82.20 (n/a)76.20 (n/a)26.31 (n/a)

dequant_1_cols_2_channels_2048_tile_1024_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.04 (-14.56%)0.04 (+0.51%)0.04 (+20.54%)0.03 (-0.86%)0.01 (-32.35%)166.20 (+0.85%)137.44 (-2.11%)126.20 (-17.03%)116.80 (+17.03%)22.00 (-20.14%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.04 (n/a)0.03 (n/a)0.03 (n/a)0.01 (n/a)164.80 (n/a)140.40 (n/a)152.10 (n/a)99.80 (n/a)27.55 (n/a)

dequant_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.07 (+5.66%)0.05 (+1.21%)0.05 (+0.61%)0.04 (-15.02%)0.01 (+58.42%)143.40 (+17.64%)101.02 (+1.49%)97.60 (-0.61%)78.90 (-5.40%)22.56 (+78.52%)
0a6c11c — 2025-12-04 00:39:100.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)121.90 (n/a)99.54 (n/a)98.20 (n/a)83.40 (n/a)12.64 (n/a)

dequant_2_cols_1_channels_2048_tile_1024_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (+65.62%)0.04 (+20.94%)0.04 (+16.97%)0.03 (-8.72%)0.01 (+413.43%)184.20 (+9.58%)135.68 (-12.18%)134.30 (-14.51%)85.70 (-39.65%)36.33 (+234.47%)
87e12ba — 2025-12-16 21:55:160.04 (n/a)0.03 (n/a)0.03 (n/a)0.03 (n/a)0.00 (n/a)168.10 (n/a)154.50 (n/a)157.10 (n/a)142.00 (n/a)10.86 (n/a)

dequant_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.05 (+17.06%)0.04 (+8.98%)0.05 (+21.37%)0.03 (-11.65%)0.01 (+301.89%)159.70 (+13.18%)126.32 (-5.68%)109.00 (-17.67%)106.80 (-14.56%)22.48 (+280.15%)
0a6c11c — 2025-12-04 00:39:100.04 (n/a)0.04 (n/a)0.04 (n/a)0.04 (n/a)0.00 (n/a)141.10 (n/a)133.92 (n/a)132.40 (n/a)125.00 (n/a)5.91 (n/a)

dequant_2_cols_2_channels_2048_tile_512_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.05 (+4.56%)0.04 (-7.03%)0.03 (-17.42%)0.02 (-27.57%)0.01 (+77.80%)211.50 (+38.05%)153.56 (+13.95%)170.20 (+21.14%)98.20 (-4.38%)45.45 (+137.43%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)153.20 (n/a)134.76 (n/a)140.50 (n/a)102.70 (n/a)19.14 (n/a)

dequant_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.06 (+20.24%)0.05 (-3.19%)0.05 (-6.73%)0.03 (-20.69%)0.01 (+194.26%)156.10 (+26.09%)120.86 (+7.15%)114.90 (+7.18%)88.30 (-16.85%)24.40 (+214.82%)
0a6c11c — 2025-12-04 00:39:100.05 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.00 (n/a)123.80 (n/a)112.80 (n/a)107.20 (n/a)106.20 (n/a)7.75 (n/a)

dequant_4_cols_1_channels_2048_tile_512_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.04 (+12.75%)0.04 (+20.57%)0.04 (+20.10%)0.03 (+43.62%)0.00 (-30.81%)153.40 (-30.40%)130.96 (-18.73%)125.10 (-16.77%)120.90 (-11.30%)13.72 (-59.25%)
87e12ba — 2025-12-16 21:55:160.04 (n/a)0.03 (n/a)0.03 (n/a)0.02 (n/a)0.01 (n/a)220.40 (n/a)161.14 (n/a)150.30 (n/a)136.30 (n/a)33.67 (n/a)

dequant_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.05 (-6.09%)0.04 (-8.04%)0.04 (-12.46%)0.03 (-0.41%)0.01 (-20.19%)168.10 (+0.42%)135.74 (+7.44%)132.20 (+14.26%)107.40 (+6.44%)22.91 (-12.67%)
0a6c11c — 2025-12-04 00:39:100.05 (n/a)0.04 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)167.40 (n/a)126.34 (n/a)115.70 (n/a)100.90 (n/a)26.23 (n/a)

dequant_4_cols_2_channels_2048_tile_256_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.04 (-2.04%)0.04 (+5.52%)0.03 (+1.37%)0.03 (+35.55%)0.00 (-60.66%)159.70 (-26.24%)149.82 (-7.78%)150.00 (-1.32%)134.00 (+2.13%)9.78 (-71.09%)
87e12ba — 2025-12-16 21:55:160.04 (n/a)0.03 (n/a)0.03 (n/a)0.02 (n/a)0.01 (n/a)216.50 (n/a)162.46 (n/a)152.00 (n/a)131.20 (n/a)33.82 (n/a)

dequant_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.04 (-13.65%)0.04 (-13.67%)0.04 (-18.56%)0.03 (-13.52%)0.00 (-23.88%)170.40 (+15.68%)140.12 (+15.33%)135.80 (+22.78%)119.60 (+15.78%)17.94 (+1.87%)
0a6c11c — 2025-12-04 00:39:100.05 (n/a)0.04 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)147.30 (n/a)121.50 (n/a)110.60 (n/a)103.30 (n/a)17.61 (n/a)

dequant_8_cols_1_channels_2048_tile_256_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.03 (-10.42%)0.03 (-0.97%)0.03 (+0.82%)0.02 (+5.32%)0.01 (-18.55%)233.50 (-5.08%)185.26 (-0.58%)160.70 (-0.80%)152.40 (+11.65%)39.99 (-16.18%)
87e12ba — 2025-12-16 21:55:160.04 (n/a)0.03 (n/a)0.03 (n/a)0.02 (n/a)0.01 (n/a)246.00 (n/a)186.34 (n/a)162.00 (n/a)136.50 (n/a)47.72 (n/a)

dequant_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.05 (+17.21%)0.03 (-1.70%)0.03 (-3.04%)0.03 (-16.20%)0.01 (+95.59%)198.20 (+19.33%)160.00 (+4.92%)166.20 (+3.17%)106.50 (-14.66%)29.86 (+87.19%)
0a6c11c — 2025-12-04 00:39:100.04 (n/a)0.03 (n/a)0.03 (n/a)0.03 (n/a)0.00 (n/a)166.10 (n/a)152.50 (n/a)161.10 (n/a)124.80 (n/a)15.95 (n/a)

dequant_8_cols_2_channels_2048_tile_128_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.04 (+28.31%)0.03 (+16.14%)0.03 (+8.41%)0.03 (+26.66%)0.00 (+23.97%)186.30 (-21.06%)165.32 (-13.99%)168.60 (-7.77%)136.00 (-22.06%)18.79 (-25.93%)
87e12ba — 2025-12-16 21:55:160.03 (n/a)0.03 (n/a)0.03 (n/a)0.02 (n/a)0.00 (n/a)236.00 (n/a)192.22 (n/a)182.80 (n/a)174.50 (n/a)25.36 (n/a)

eltwise_add_1_cols_2_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.11 (+2.09%)0.09 (+0.38%)0.09 (-5.16%)0.08 (+24.05%)0.01 (-34.62%)158.20 (-19.41%)137.36 (-2.86%)132.40 (+5.50%)113.80 (-2.07%)17.63 (-47.67%)
87e12ba — 2025-12-16 21:55:160.11 (n/a)0.09 (n/a)0.10 (n/a)0.06 (n/a)0.02 (n/a)196.30 (n/a)141.40 (n/a)125.50 (n/a)116.20 (n/a)33.70 (n/a)

eltwise_add_2_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.10 (+3.70%)0.09 (+5.63%)0.09 (+11.82%)0.07 (+5.59%)0.01 (-5.51%)170.90 (-5.32%)138.76 (-5.61%)132.30 (-10.55%)118.70 (-3.50%)20.04 (-11.80%)
87e12ba — 2025-12-16 21:55:160.10 (n/a)0.09 (n/a)0.08 (n/a)0.07 (n/a)0.01 (n/a)180.50 (n/a)147.00 (n/a)147.90 (n/a)123.00 (n/a)22.72 (n/a)

eltwise_add_4_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.11 (+21.93%)0.10 (+33.29%)0.10 (+45.64%)0.09 (+25.11%)0.01 (+4.51%)143.60 (-20.04%)122.04 (-25.25%)118.60 (-31.33%)111.00 (-18.02%)13.46 (-32.64%)
87e12ba — 2025-12-16 21:55:160.09 (n/a)0.08 (n/a)0.07 (n/a)0.07 (n/a)0.01 (n/a)179.60 (n/a)163.26 (n/a)172.70 (n/a)135.40 (n/a)19.98 (n/a)

eltwise_add_8_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.11 (+24.58%)0.09 (+29.81%)0.09 (+38.16%)0.07 (+47.26%)0.01 (-15.80%)169.40 (-32.08%)141.64 (-25.33%)141.30 (-27.61%)109.30 (-19.69%)21.51 (-54.39%)
87e12ba — 2025-12-16 21:55:160.09 (n/a)0.07 (n/a)0.06 (n/a)0.05 (n/a)0.02 (n/a)249.40 (n/a)189.68 (n/a)195.20 (n/a)136.10 (n/a)47.16 (n/a)

eltwise_mul_1_cols_2_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.11 (+4.41%)0.09 (+4.03%)0.09 (+12.65%)0.08 (+11.17%)0.01 (-18.95%)150.80 (-10.02%)135.40 (-4.59%)131.80 (-11.25%)113.40 (-4.22%)15.51 (-27.31%)
87e12ba — 2025-12-16 21:55:160.10 (n/a)0.09 (n/a)0.08 (n/a)0.07 (n/a)0.01 (n/a)167.60 (n/a)141.92 (n/a)148.50 (n/a)118.40 (n/a)21.34 (n/a)

eltwise_mul_2_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.11 (+14.90%)0.09 (+8.15%)0.09 (+4.02%)0.08 (+5.15%)0.01 (+55.82%)147.30 (-4.91%)132.16 (-6.97%)134.60 (-3.86%)109.50 (-13.03%)16.05 (+28.59%)
87e12ba — 2025-12-16 21:55:160.10 (n/a)0.09 (n/a)0.09 (n/a)0.08 (n/a)0.01 (n/a)154.90 (n/a)142.06 (n/a)140.00 (n/a)125.90 (n/a)12.48 (n/a)

eltwise_mul_4_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.10 (-1.25%)0.09 (+4.14%)0.08 (+2.09%)0.08 (+32.93%)0.01 (-49.07%)157.60 (-24.74%)145.06 (-6.64%)149.80 (-2.09%)123.90 (+1.23%)13.01 (-61.81%)
87e12ba — 2025-12-16 21:55:160.10 (n/a)0.08 (n/a)0.08 (n/a)0.06 (n/a)0.02 (n/a)209.40 (n/a)155.38 (n/a)153.00 (n/a)122.40 (n/a)34.08 (n/a)

eltwise_mul_8_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.10 (-5.34%)0.08 (-2.83%)0.08 (-8.04%)0.06 (+5.07%)0.01 (-22.63%)201.10 (-4.87%)155.80 (+1.30%)149.50 (+8.73%)124.30 (+5.61%)28.24 (-22.32%)
87e12ba — 2025-12-16 21:55:160.10 (n/a)0.08 (n/a)0.09 (n/a)0.06 (n/a)0.02 (n/a)211.40 (n/a)153.80 (n/a)137.50 (n/a)117.70 (n/a)36.35 (n/a)

gelu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-6.46%)0.06 (+0.08%)0.06 (-1.55%)0.05 (+3.12%)0.01 (-20.66%)157.50 (-3.02%)139.32 (-0.56%)143.30 (+1.56%)119.60 (+6.88%)15.25 (-16.11%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)162.40 (n/a)140.10 (n/a)141.10 (n/a)111.90 (n/a)18.18 (n/a)

gelu_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.10 (+54.19%)0.07 (+28.24%)0.07 (+29.84%)0.04 (+0.93%)0.02 (+136.31%)222.60 (-0.93%)142.18 (-14.59%)114.30 (-23.03%)84.30 (-35.15%)58.47 (+54.36%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.05 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)224.70 (n/a)166.46 (n/a)148.50 (n/a)130.00 (n/a)37.88 (n/a)

gelu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-12.40%)0.06 (+3.39%)0.06 (+7.39%)0.05 (+11.66%)0.01 (-44.52%)155.70 (-10.41%)133.88 (-5.56%)130.20 (-6.93%)113.60 (+14.17%)16.28 (-41.46%)
87e12ba — 2025-12-16 21:55:160.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)173.80 (n/a)141.76 (n/a)139.90 (n/a)99.50 (n/a)27.81 (n/a)

gelu_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.08 (+13.52%)0.06 (+8.99%)0.06 (+13.63%)0.04 (+0.59%)0.01 (+30.90%)210.80 (-0.57%)148.52 (-6.69%)140.50 (-11.97%)107.20 (-11.91%)39.27 (+15.79%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)212.00 (n/a)159.16 (n/a)159.60 (n/a)121.70 (n/a)33.92 (n/a)

gelu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+6.91%)0.06 (-0.98%)0.06 (+0.12%)0.04 (-21.04%)0.01 (+85.60%)204.80 (+26.65%)150.04 (+3.93%)144.10 (-0.07%)115.10 (-6.50%)34.48 (+122.89%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)161.70 (n/a)144.36 (n/a)144.20 (n/a)123.10 (n/a)15.47 (n/a)

gelu_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+48.20%)0.05 (+25.60%)0.05 (+18.65%)0.04 (+5.75%)0.01 (+178.71%)223.30 (-5.46%)161.54 (-16.57%)155.90 (-15.73%)114.70 (-32.53%)44.56 (+72.96%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)236.20 (n/a)193.62 (n/a)185.00 (n/a)170.00 (n/a)25.77 (n/a)

gelu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+3.15%)0.06 (+11.97%)0.06 (+12.83%)0.04 (+14.93%)0.01 (-10.01%)187.30 (-13.01%)142.26 (-11.72%)128.10 (-11.41%)122.90 (-3.08%)26.88 (-24.65%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.05 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)215.30 (n/a)161.14 (n/a)144.60 (n/a)126.80 (n/a)35.68 (n/a)

gelu_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (+19.55%)0.05 (+19.02%)0.06 (+30.50%)0.03 (-0.75%)0.01 (+73.54%)236.70 (+0.77%)169.76 (-13.83%)148.40 (-23.35%)135.00 (-16.36%)41.76 (+47.23%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)234.90 (n/a)197.00 (n/a)193.60 (n/a)161.40 (n/a)28.37 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.72 (-0.11%)3.55 (-2.75%)3.56 (-2.75%)3.36 (-5.33%)0.12 (+76.10%)2725.23 (-0.11%)2602.52 (-2.75%)2612.11 (-2.75%)2459.89 (-5.33%)86.27 (+76.10%)6984.00 (+5.63%)6608.60 (+2.91%)6577.00 (+2.83%)6304.00 (+0.11%)221.87 (+87.15%)
9e98049 — 2025-11-25 20:50:553.72 (n/a)3.65 (n/a)3.66 (n/a)3.54 (n/a)0.07 (n/a)2728.26 (n/a)2676.06 (n/a)2686.03 (n/a)2598.29 (n/a)48.99 (n/a)6612.00 (n/a)6422.00 (n/a)6396.00 (n/a)6297.00 (n/a)118.55 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.40 (-3.14%)3.38 (-0.82%)3.38 (-0.82%)3.35 (+4.93%)0.02 (-84.51%)2493.81 (-3.14%)2475.27 (-0.82%)2479.06 (-0.82%)2453.92 (+4.93%)13.33 (-84.51%)7001.00 (-4.70%)6940.80 (+0.71%)6930.00 (+0.83%)6889.00 (+3.24%)37.43 (-84.76%)
9e98049 — 2025-11-25 20:50:553.51 (n/a)3.40 (n/a)3.41 (n/a)3.19 (n/a)0.12 (n/a)2574.53 (n/a)2495.72 (n/a)2499.62 (n/a)2338.67 (n/a)86.02 (n/a)7346.00 (n/a)6892.20 (n/a)6873.00 (n/a)6673.00 (n/a)245.57 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.02 (+0.52%)2.87 (-2.05%)2.86 (-2.24%)2.70 (-6.16%)0.11 (+131.66%)2216.19 (+0.52%)2104.44 (-2.05%)2094.08 (-2.24%)1980.84 (-6.16%)80.25 (+131.67%)8673.00 (+6.56%)8175.60 (+2.21%)8204.00 (+2.29%)7752.00 (-0.51%)314.03 (+145.36%)
9e98049 — 2025-11-25 20:50:553.01 (n/a)2.93 (n/a)2.92 (n/a)2.88 (n/a)0.05 (n/a)2204.81 (n/a)2148.41 (n/a)2142.13 (n/a)2110.81 (n/a)34.64 (n/a)8139.00 (n/a)7998.60 (n/a)8020.00 (n/a)7792.00 (n/a)127.99 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:491.04 (+0.94%)1.03 (+0.23%)1.03 (+0.09%)1.02 (+0.07%)0.00 (+197.51%)24576.00 (-0.07%)24485.82 (-0.23%)24530.40 (-0.09%)24266.40 (-0.94%)111.26 (+194.35%)
0a6c11c — 2025-12-04 00:39:101.03 (n/a)1.03 (n/a)1.03 (n/a)1.02 (n/a)0.00 (n/a)24594.30 (n/a)24542.14 (n/a)24551.90 (n/a)24495.50 (n/a)37.80 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:331.03 (+0.05%)1.02 (-0.08%)1.03 (+0.07%)1.02 (-0.31%)0.00 (+563.47%)24626.00 (+0.31%)24560.44 (+0.08%)24523.80 (-0.07%)24517.00 (-0.05%)54.29 (+563.95%)700.73 (+0.05%)699.50 (-0.08%)700.54 (+0.07%)697.63 (-0.31%)1.55 (+563.96%)
87e12ba — 2025-12-16 21:55:161.03 (n/a)1.03 (n/a)1.03 (n/a)1.03 (n/a)0.00 (n/a)24549.80 (n/a)24541.26 (n/a)24541.70 (n/a)24530.00 (n/a)8.18 (n/a)700.36 (n/a)700.04 (n/a)700.03 (n/a)699.80 (n/a)0.23 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.99 (+1.69%)0.97 (+0.29%)0.97 (-0.04%)0.97 (-0.15%)0.01 (+434.92%)26025.30 (+0.15%)25839.34 (-0.29%)25916.70 (+0.04%)25446.00 (-1.67%)202.05 (+425.44%)
0a6c11c — 2025-12-04 00:39:100.97 (n/a)0.97 (n/a)0.97 (n/a)0.97 (n/a)0.00 (n/a)25986.90 (n/a)25913.52 (n/a)25905.70 (n/a)25877.00 (n/a)38.45 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:330.97 (+0.02%)0.97 (-0.19%)0.97 (-0.04%)0.97 (-0.57%)0.00 (+717.52%)26054.70 (+0.57%)25943.22 (+0.19%)25904.50 (+0.04%)25879.60 (-0.02%)78.16 (+723.45%)663.84 (+0.02%)662.21 (-0.19%)663.20 (-0.04%)659.38 (-0.57%)1.99 (+717.54%)
87e12ba — 2025-12-16 21:55:160.97 (n/a)0.97 (n/a)0.97 (n/a)0.97 (n/a)0.00 (n/a)25905.80 (n/a)25893.90 (n/a)25893.20 (n/a)25884.30 (n/a)9.49 (n/a)663.72 (n/a)663.47 (n/a)663.49 (n/a)663.17 (n/a)0.24 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.92 (-1.02%)0.92 (-0.66%)0.92 (-0.10%)0.91 (-0.78%)0.01 (-4.54%)27751.40 (+0.79%)27480.74 (+0.66%)27354.80 (+0.10%)27320.50 (+1.03%)182.69 (-2.84%)
0a6c11c — 2025-12-04 00:39:100.93 (n/a)0.92 (n/a)0.92 (n/a)0.91 (n/a)0.01 (n/a)27534.80 (n/a)27299.32 (n/a)27326.70 (n/a)27042.30 (n/a)188.04 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:330.92 (-0.43%)0.91 (-0.02%)0.91 (+0.21%)0.91 (+0.16%)0.00 (-14.29%)27740.30 (-0.16%)27586.74 (+0.02%)27504.50 (-0.21%)27476.50 (+0.44%)134.04 (-14.02%)625.26 (-0.43%)622.77 (-0.02%)624.62 (+0.21%)619.31 (+0.16%)3.02 (-14.29%)
87e12ba — 2025-12-16 21:55:160.92 (n/a)0.91 (n/a)0.91 (n/a)0.91 (n/a)0.01 (n/a)27785.00 (n/a)27580.62 (n/a)27563.20 (n/a)27357.10 (n/a)155.90 (n/a)627.99 (n/a)622.91 (n/a)623.29 (n/a)618.32 (n/a)3.52 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
31ab24e — 2025-11-22 02:01:563.65 (-1.06%)3.59 (+0.83%)3.63 (-0.26%)3.45 (+2.66%)0.07 (-44.71%)2678.08 (-1.06%)2632.13 (+0.83%)2658.60 (-0.26%)2528.31 (+2.66%)53.86 (-44.71%)6795.00 (-2.59%)6529.80 (-0.92%)6462.00 (+0.26%)6415.00 (+1.07%)137.14 (-45.27%)
d9a64c5 — 2025-11-21 20:49:353.69 (n/a)3.56 (n/a)3.64 (n/a)3.36 (n/a)0.13 (n/a)2706.77 (n/a)2610.50 (n/a)2665.61 (n/a)2462.71 (n/a)97.41 (n/a)6976.00 (n/a)6590.40 (n/a)6445.00 (n/a)6347.00 (n/a)250.58 (n/a)

layer_norm_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+1.68%)0.06 (-0.78%)0.06 (-3.61%)0.05 (+0.29%)0.01 (+20.71%)172.30 (-0.29%)141.66 (+1.68%)141.50 (+3.74%)113.30 (-1.65%)27.00 (+17.49%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)172.80 (n/a)139.32 (n/a)136.40 (n/a)115.20 (n/a)22.98 (n/a)

layer_norm_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+4.51%)0.06 (+1.80%)0.06 (-5.71%)0.05 (+24.88%)0.01 (-21.66%)172.10 (-19.92%)146.22 (-3.83%)148.30 (+6.08%)114.90 (-4.33%)20.42 (-44.23%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)214.90 (n/a)152.04 (n/a)139.80 (n/a)120.10 (n/a)36.62 (n/a)

layer_norm_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.08 (+15.12%)0.06 (+7.50%)0.06 (+9.24%)0.05 (-11.75%)0.01 (+87.93%)178.30 (+13.35%)138.52 (-5.27%)134.80 (-8.49%)107.70 (-13.15%)25.33 (+88.89%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)157.30 (n/a)146.22 (n/a)147.30 (n/a)124.00 (n/a)13.41 (n/a)

layer_norm_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-0.33%)0.06 (+10.15%)0.06 (+5.77%)0.05 (+54.23%)0.01 (-40.26%)156.50 (-35.14%)142.22 (-13.21%)147.80 (-5.44%)113.60 (+0.35%)17.74 (-62.47%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.05 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)241.30 (n/a)163.86 (n/a)156.30 (n/a)113.20 (n/a)47.25 (n/a)

layer_norm_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-4.63%)0.05 (+2.22%)0.05 (-2.84%)0.05 (+36.10%)0.00 (-61.42%)172.80 (-26.50%)156.30 (-5.71%)153.00 (+2.89%)140.70 (+4.84%)12.24 (-70.47%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.05 (n/a)0.06 (n/a)0.03 (n/a)0.01 (n/a)235.10 (n/a)165.76 (n/a)148.70 (n/a)134.20 (n/a)41.44 (n/a)

layer_norm_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (+31.39%)0.05 (+18.06%)0.05 (+6.45%)0.05 (+20.91%)0.01 (+32.19%)173.70 (-17.29%)158.22 (-15.25%)163.00 (-6.11%)128.90 (-23.91%)17.11 (-19.57%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.04 (n/a)0.05 (n/a)0.04 (n/a)0.00 (n/a)210.00 (n/a)186.70 (n/a)173.60 (n/a)169.40 (n/a)21.28 (n/a)

layer_norm_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-10.74%)0.05 (-6.45%)0.05 (+4.42%)0.05 (-4.80%)0.01 (-22.06%)181.50 (+5.03%)157.66 (+6.34%)151.40 (-4.24%)133.40 (+12.10%)21.64 (-5.54%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)172.80 (n/a)148.26 (n/a)158.10 (n/a)119.00 (n/a)22.91 (n/a)

layer_norm_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-0.52%)0.05 (+10.03%)0.06 (+5.99%)0.05 (+40.18%)0.00 (-50.41%)177.10 (-28.67%)154.66 (-12.09%)146.10 (-5.62%)142.00 (+0.50%)15.01 (-65.28%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.05 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)248.30 (n/a)175.92 (n/a)154.80 (n/a)141.30 (n/a)43.24 (n/a)

matrix_vector_mul_128x128_32_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:490.37 (+17.13%)0.30 (+24.87%)0.30 (+38.03%)0.23 (+12.47%)0.06 (+35.23%)
0a6c11c — 2025-12-04 00:39:100.32 (n/a)0.24 (n/a)0.22 (n/a)0.20 (n/a)0.04 (n/a)

matrix_vector_mul_128x128_32_1col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:330.30 (-0.32%)0.24 (-6.08%)0.22 (-6.96%)0.21 (-3.63%)0.04 (-2.67%)0.29 (-0.32%)0.24 (-6.08%)0.22 (-6.96%)0.21 (-3.63%)0.03 (-2.67%)
87e12ba — 2025-12-16 21:55:160.30 (n/a)0.26 (n/a)0.24 (n/a)0.22 (n/a)0.04 (n/a)0.29 (n/a)0.25 (n/a)0.24 (n/a)0.21 (n/a)0.04 (n/a)

matrix_vector_mul_2048x8192_1_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4913.57 (+3.47%)13.15 (+1.09%)13.10 (+0.16%)12.52 (-1.58%)0.37 (+153.19%)
0a6c11c — 2025-12-04 00:39:1013.11 (n/a)13.01 (n/a)13.07 (n/a)12.72 (n/a)0.15 (n/a)

matrix_vector_mul_2048x8192_1_1col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:3313.34 (+0.27%)13.06 (+4.10%)13.01 (-1.36%)12.72 (+23.40%)0.27 (-79.10%)13.34 (+0.27%)13.05 (+4.10%)13.00 (-1.36%)12.71 (+23.40%)0.27 (-79.10%)
87e12ba — 2025-12-16 21:55:1613.31 (n/a)12.54 (n/a)13.19 (n/a)10.31 (n/a)1.28 (n/a)13.30 (n/a)12.54 (n/a)13.18 (n/a)10.30 (n/a)1.28 (n/a)

matrix_vector_mul_2048x8192_1_2col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4925.06 (+2.47%)23.71 (+1.81%)24.13 (+0.50%)20.90 (+5.41%)1.48 (-15.60%)
0a6c11c — 2025-12-04 00:39:1024.46 (n/a)23.29 (n/a)24.01 (n/a)19.83 (n/a)1.75 (n/a)

matrix_vector_mul_2048x8192_1_2col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:3325.00 (+2.71%)24.28 (+8.92%)24.61 (+2.44%)22.83 (+27.49%)0.91 (-67.63%)24.98 (+2.71%)24.26 (+8.92%)24.60 (+2.44%)22.81 (+27.49%)0.91 (-67.63%)
87e12ba — 2025-12-16 21:55:1624.34 (n/a)22.29 (n/a)24.03 (n/a)17.90 (n/a)2.80 (n/a)24.32 (n/a)22.27 (n/a)24.01 (n/a)17.89 (n/a)2.80 (n/a)

matrix_vector_mul_2048x8192_1_4col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4942.05 (+1.54%)40.50 (+12.60%)40.49 (+2.36%)38.98 (+97.53%)0.97 (-88.09%)
0a6c11c — 2025-12-04 00:39:1041.41 (n/a)35.97 (n/a)39.55 (n/a)19.73 (n/a)8.15 (n/a)

matrix_vector_mul_2048x8192_1_4col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:3339.58 (-0.60%)39.02 (-0.81%)39.26 (-0.75%)38.35 (-0.22%)0.56 (+4.28%)39.55 (-0.60%)39.00 (-0.81%)39.23 (-0.75%)38.32 (-0.22%)0.56 (+4.28%)
87e12ba — 2025-12-16 21:55:1639.82 (n/a)39.34 (n/a)39.55 (n/a)38.43 (n/a)0.54 (n/a)39.79 (n/a)39.32 (n/a)39.53 (n/a)38.41 (n/a)0.54 (n/a)

matrix_vector_mul_2048x8192_1_8col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4946.62 (+3.11%)44.82 (+0.17%)44.78 (+0.17%)42.60 (-3.80%)1.60 (+367.72%)
0a6c11c — 2025-12-04 00:39:1045.21 (n/a)44.75 (n/a)44.70 (n/a)44.29 (n/a)0.34 (n/a)

matrix_vector_mul_2048x8192_1_8col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:3346.30 (+5.10%)40.99 (-3.38%)44.22 (+2.70%)24.65 (-36.58%)9.18 (+334.14%)46.27 (+5.10%)40.96 (-3.38%)44.19 (+2.70%)24.63 (-36.58%)9.18 (+334.14%)
87e12ba — 2025-12-16 21:55:1644.05 (n/a)42.42 (n/a)43.06 (n/a)38.86 (n/a)2.11 (n/a)44.03 (n/a)42.39 (n/a)43.03 (n/a)38.84 (n/a)2.11 (n/a)

matrix_vector_mul_8192x2048_4_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4913.29 (+1.63%)13.19 (+6.08%)13.19 (+1.17%)13.09 (+26.86%)0.08 (-92.94%)
0a6c11c — 2025-12-04 00:39:1013.08 (n/a)12.44 (n/a)13.03 (n/a)10.32 (n/a)1.07 (n/a)

matrix_vector_mul_8192x2048_4_1col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:3313.45 (+0.47%)12.94 (-0.64%)13.27 (+1.57%)11.34 (-10.16%)0.90 (+150.53%)13.45 (+0.47%)12.93 (-0.64%)13.26 (+1.57%)11.33 (-10.16%)0.90 (+150.53%)
87e12ba — 2025-12-16 21:55:1613.39 (n/a)13.02 (n/a)13.06 (n/a)12.62 (n/a)0.36 (n/a)13.38 (n/a)13.01 (n/a)13.06 (n/a)12.62 (n/a)0.36 (n/a)

matrix_vector_mul_8192x2048_4_2col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4925.58 (+4.24%)24.23 (+6.59%)24.03 (+0.29%)23.58 (+37.29%)0.70 (-74.97%)
0a6c11c — 2025-12-04 00:39:1024.54 (n/a)22.73 (n/a)23.96 (n/a)17.18 (n/a)2.79 (n/a)

matrix_vector_mul_8192x2048_4_2col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:3324.50 (-3.07%)22.66 (-5.24%)24.10 (-1.53%)17.55 (-14.73%)2.93 (+54.50%)24.48 (-3.07%)22.65 (-5.24%)24.09 (-1.53%)17.54 (-14.73%)2.93 (+54.50%)
87e12ba — 2025-12-16 21:55:1625.28 (n/a)23.92 (n/a)24.48 (n/a)20.58 (n/a)1.90 (n/a)25.26 (n/a)23.90 (n/a)24.46 (n/a)20.57 (n/a)1.90 (n/a)

matrix_vector_mul_8192x2048_4_4col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4940.94 (-2.19%)39.74 (-1.10%)39.76 (-0.60%)38.17 (-2.14%)1.04 (+4.39%)
0a6c11c — 2025-12-04 00:39:1041.86 (n/a)40.18 (n/a)40.00 (n/a)39.00 (n/a)1.00 (n/a)

matrix_vector_mul_8192x2048_4_4col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:3341.03 (+3.50%)39.21 (+0.89%)39.01 (+0.08%)38.18 (+1.06%)1.07 (+58.98%)41.00 (+3.50%)39.18 (+0.89%)38.99 (+0.08%)38.16 (+1.06%)1.07 (+58.98%)
87e12ba — 2025-12-16 21:55:1639.64 (n/a)38.86 (n/a)38.98 (n/a)37.78 (n/a)0.67 (n/a)39.61 (n/a)38.84 (n/a)38.95 (n/a)37.76 (n/a)0.67 (n/a)

matrix_vector_mul_8192x2048_4_8col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4946.71 (-1.04%)44.91 (-2.34%)44.71 (-2.98%)43.80 (-1.44%)0.96 (+6.93%)
0a6c11c — 2025-12-04 00:39:1047.20 (n/a)45.98 (n/a)46.08 (n/a)44.44 (n/a)0.90 (n/a)

matrix_vector_mul_8192x2048_4_8col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
68acccf — 2025-12-16 23:07:3344.45 (+0.29%)40.43 (-7.31%)43.33 (-0.87%)27.32 (-35.93%)7.34 (+1073.29%)44.43 (+0.29%)40.41 (-7.31%)43.30 (-0.87%)27.30 (-35.93%)7.34 (+1073.29%)
87e12ba — 2025-12-16 21:55:1644.32 (n/a)43.62 (n/a)43.71 (n/a)42.64 (n/a)0.63 (n/a)44.30 (n/a)43.59 (n/a)43.68 (n/a)42.62 (n/a)0.63 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.07 (+22.03%)0.06 (+27.86%)0.06 (+35.49%)0.05 (+15.00%)0.01 (+49.69%)173.30 (-13.00%)137.94 (-21.32%)133.60 (-26.19%)123.40 (-18.06%)18.40 (+8.53%)
0a6c11c — 2025-12-04 00:39:100.05 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.00 (n/a)199.20 (n/a)175.32 (n/a)181.00 (n/a)150.60 (n/a)16.96 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (+1.33%)0.05 (+12.74%)0.06 (+17.09%)0.04 (-1.39%)0.01 (+13.39%)214.90 (+1.42%)157.32 (-10.72%)145.30 (-14.63%)137.00 (-1.30%)32.81 (+14.59%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)211.90 (n/a)176.20 (n/a)170.20 (n/a)138.80 (n/a)28.64 (n/a)

mem_copy_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (-16.99%)0.05 (-12.47%)0.05 (-7.32%)0.04 (-11.17%)0.00 (-29.61%)197.00 (+12.57%)171.00 (+13.70%)164.00 (+7.89%)153.00 (+20.47%)17.49 (-4.32%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)175.00 (n/a)150.40 (n/a)152.00 (n/a)127.00 (n/a)18.28 (n/a)

mem_copy_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (+5.56%)0.05 (-1.59%)0.05 (-1.39%)0.03 (-14.07%)0.01 (+25.03%)263.00 (+16.37%)166.00 (+4.53%)144.00 (+1.41%)126.00 (-5.26%)49.38 (+43.17%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)226.00 (n/a)158.80 (n/a)142.00 (n/a)133.00 (n/a)34.49 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.08 (+10.93%)0.07 (+7.79%)0.07 (+3.55%)0.05 (-1.73%)0.01 (+36.73%)150.20 (+1.76%)121.68 (-6.36%)125.60 (-3.38%)97.90 (-9.85%)18.56 (+21.98%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.06 (n/a)0.06 (n/a)0.06 (n/a)0.01 (n/a)147.60 (n/a)129.94 (n/a)130.00 (n/a)108.60 (n/a)15.21 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-14.92%)0.06 (-3.23%)0.06 (+11.58%)0.04 (-15.32%)0.01 (-15.39%)188.30 (+18.13%)139.12 (+3.39%)126.00 (-10.38%)121.90 (+17.55%)27.94 (+19.68%)
87e12ba — 2025-12-16 21:55:160.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)159.40 (n/a)134.56 (n/a)140.60 (n/a)103.70 (n/a)23.35 (n/a)

mem_copy_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (-15.32%)0.05 (-17.09%)0.05 (-23.81%)0.04 (-15.54%)0.01 (-18.18%)193.00 (+18.40%)160.80 (+20.36%)168.00 (+31.25%)124.00 (+18.10%)22.92 (+11.39%)
0999a9a — 2025-11-13 19:24:080.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)163.00 (n/a)133.60 (n/a)128.00 (n/a)105.00 (n/a)20.58 (n/a)

mem_copy_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (-23.02%)0.05 (-6.98%)0.05 (-2.13%)0.04 (-15.31%)0.01 (-33.29%)196.00 (+18.07%)147.00 (+6.52%)141.00 (+2.17%)126.00 (+29.90%)25.27 (+8.13%)
0999a9a — 2025-11-13 19:24:080.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)166.00 (n/a)138.00 (n/a)138.00 (n/a)97.00 (n/a)23.37 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.10 (+31.37%)0.08 (+29.12%)0.08 (+29.06%)0.07 (+41.80%)0.01 (+0.23%)109.90 (-29.46%)98.30 (-23.38%)100.80 (-22.52%)79.40 (-23.87%)10.54 (-46.17%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.07 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)155.80 (n/a)128.30 (n/a)130.10 (n/a)104.30 (n/a)19.58 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-5.20%)0.06 (+0.26%)0.07 (+12.21%)0.05 (-11.65%)0.01 (+39.39%)180.80 (+13.21%)141.30 (+1.65%)125.10 (-10.90%)117.30 (+5.49%)29.86 (+67.65%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)159.70 (n/a)139.00 (n/a)140.40 (n/a)111.20 (n/a)17.81 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.10 (+25.02%)0.07 (+10.01%)0.07 (+7.53%)0.05 (-1.99%)0.01 (+77.08%)150.10 (+2.04%)114.60 (-7.22%)111.70 (-6.99%)83.90 (-20.02%)21.23 (+43.54%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)147.10 (n/a)123.52 (n/a)120.10 (n/a)104.90 (n/a)14.79 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.08 (+13.53%)0.07 (+24.45%)0.07 (+28.38%)0.06 (+52.79%)0.01 (-24.75%)148.80 (-34.56%)122.38 (-22.50%)120.40 (-22.12%)98.10 (-11.94%)18.10 (-57.85%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)227.40 (n/a)157.90 (n/a)154.60 (n/a)111.40 (n/a)42.94 (n/a)

mem_copy_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.07 (+11.97%)0.05 (+3.09%)0.05 (+2.10%)0.03 (-11.35%)0.01 (+41.65%)229.00 (+12.81%)156.00 (-0.51%)143.00 (-2.05%)117.00 (-10.69%)38.21 (+48.71%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)203.00 (n/a)156.80 (n/a)146.00 (n/a)131.00 (n/a)25.69 (n/a)

mem_copy_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (+9.02%)0.05 (+5.59%)0.05 (+0.00%)0.04 (+3.63%)0.01 (+25.48%)193.00 (-3.50%)147.40 (-4.66%)147.00 (+0.00%)122.00 (-8.27%)25.24 (+7.48%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)200.00 (n/a)154.60 (n/a)147.00 (n/a)133.00 (n/a)23.48 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.09 (-3.98%)0.07 (+4.00%)0.07 (+7.04%)0.06 (+3.67%)0.01 (-15.76%)129.70 (-3.57%)113.30 (-4.50%)116.90 (-6.55%)90.40 (+4.15%)15.04 (-13.80%)
0a6c11c — 2025-12-04 00:39:100.09 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)134.50 (n/a)118.64 (n/a)125.10 (n/a)86.80 (n/a)17.45 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.09 (+34.81%)0.06 (+4.99%)0.06 (+0.65%)0.04 (-21.47%)0.02 (+288.64%)196.20 (+27.40%)137.34 (+0.31%)131.60 (-0.60%)95.00 (-25.84%)37.44 (+268.84%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.00 (n/a)154.00 (n/a)136.92 (n/a)132.40 (n/a)128.10 (n/a)10.15 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.08 (+9.57%)0.07 (+16.85%)0.07 (+19.58%)0.06 (+24.52%)0.01 (+1.05%)134.50 (-19.70%)116.64 (-14.92%)116.80 (-16.39%)98.00 (-8.75%)15.71 (-25.13%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)167.50 (n/a)137.10 (n/a)139.70 (n/a)107.40 (n/a)20.99 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.08 (+31.03%)0.07 (+22.90%)0.07 (+28.95%)0.05 (+7.71%)0.01 (+133.55%)152.80 (-7.11%)122.52 (-17.31%)115.20 (-22.42%)99.80 (-23.64%)20.58 (+68.42%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.00 (n/a)164.50 (n/a)148.16 (n/a)148.50 (n/a)130.70 (n/a)12.22 (n/a)

mem_copy_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (-13.48%)0.05 (-17.79%)0.05 (-15.33%)0.03 (-25.09%)0.01 (+6.67%)267.00 (+33.50%)176.80 (+24.51%)150.00 (+18.11%)141.00 (+15.57%)47.41 (+61.41%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)200.00 (n/a)142.00 (n/a)127.00 (n/a)122.00 (n/a)29.37 (n/a)

mem_copy_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (+13.38%)0.05 (+23.85%)0.05 (+37.91%)0.05 (+30.72%)0.00 (-56.85%)166.00 (-23.50%)153.40 (-20.60%)153.00 (-27.49%)142.00 (-11.80%)7.63 (-70.60%)
0999a9a — 2025-11-13 19:24:080.05 (n/a)0.04 (n/a)0.04 (n/a)0.04 (n/a)0.01 (n/a)217.00 (n/a)193.20 (n/a)211.00 (n/a)161.00 (n/a)25.96 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.09 (+16.31%)0.08 (+15.34%)0.08 (+14.60%)0.07 (+19.30%)0.01 (-5.41%)117.30 (-16.15%)107.80 (-13.61%)108.50 (-12.78%)94.10 (-13.99%)8.29 (-32.01%)
0a6c11c — 2025-12-04 00:39:100.07 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)139.90 (n/a)124.78 (n/a)124.40 (n/a)109.40 (n/a)12.20 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.08 (+10.60%)0.07 (+8.27%)0.07 (+13.26%)0.05 (+7.26%)0.02 (+34.24%)173.60 (-6.77%)128.18 (-6.12%)112.50 (-11.70%)98.90 (-9.60%)33.17 (+10.13%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)186.20 (n/a)136.54 (n/a)127.40 (n/a)109.40 (n/a)30.12 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.09 (+12.61%)0.07 (-2.24%)0.07 (-8.85%)0.06 (+5.13%)0.01 (+35.50%)136.50 (-4.88%)120.86 (+2.96%)125.10 (+9.64%)91.60 (-11.15%)15.72 (+10.46%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)143.50 (n/a)117.38 (n/a)114.10 (n/a)103.10 (n/a)14.23 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-6.42%)0.06 (+5.73%)0.06 (+10.39%)0.05 (+8.34%)0.00 (-42.91%)157.40 (-7.68%)139.88 (-6.47%)138.40 (-9.42%)125.60 (+6.89%)11.74 (-42.53%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)170.50 (n/a)149.56 (n/a)152.80 (n/a)117.50 (n/a)20.43 (n/a)

mha

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.18 (+0.16%)0.18 (+0.19%)0.18 (+0.12%)0.18 (+0.32%)0.00 (-52.24%)47253.90 (-0.32%)47213.70 (-0.19%)47210.90 (-0.12%)47167.70 (-0.16%)28.71 (-52.47%)
0a6c11c — 2025-12-04 00:39:100.18 (n/a)0.18 (n/a)0.18 (n/a)0.18 (n/a)0.00 (n/a)47405.00 (n/a)47301.72 (n/a)47265.90 (n/a)47241.80 (n/a)60.40 (n/a)

mha0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.18 (-0.02%)0.18 (+0.04%)0.18 (+0.10%)0.18 (+0.03%)0.00 (-8.98%)47282.50 (-0.03%)47191.42 (-0.04%)47237.70 (-0.10%)46959.00 (+0.02%)133.92 (-9.02%)
87e12ba — 2025-12-16 21:55:160.18 (n/a)0.18 (n/a)0.18 (n/a)0.18 (n/a)0.00 (n/a)47298.40 (n/a)47210.22 (n/a)47285.50 (n/a)46951.70 (n/a)147.19 (n/a)

relu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+8.74%)0.06 (+4.18%)0.06 (-1.52%)0.05 (+1.48%)0.01 (+23.91%)173.70 (-1.42%)144.06 (-3.55%)145.30 (+1.54%)115.20 (-8.06%)21.44 (+11.11%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)176.20 (n/a)149.36 (n/a)143.10 (n/a)125.30 (n/a)19.29 (n/a)

relu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-6.57%)0.06 (-5.33%)0.05 (-12.18%)0.05 (+4.22%)0.01 (-13.32%)164.80 (-4.02%)149.62 (+5.28%)161.20 (+13.92%)122.90 (+7.06%)18.45 (-9.84%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)171.70 (n/a)142.12 (n/a)141.50 (n/a)114.80 (n/a)20.46 (n/a)

relu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-21.53%)0.05 (-14.15%)0.05 (-14.02%)0.05 (-7.93%)0.00 (-50.04%)170.00 (+8.63%)152.06 (+15.11%)155.50 (+16.31%)138.30 (+27.35%)13.38 (-31.51%)
87e12ba — 2025-12-16 21:55:160.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)156.50 (n/a)132.10 (n/a)133.70 (n/a)108.60 (n/a)19.53 (n/a)

relu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-0.95%)0.06 (+1.76%)0.06 (-2.28%)0.05 (-6.83%)0.01 (+11.26%)178.90 (+7.32%)145.86 (-1.42%)143.10 (+2.29%)128.70 (+0.94%)20.62 (+16.13%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)166.70 (n/a)147.96 (n/a)139.90 (n/a)127.50 (n/a)17.75 (n/a)

rms_norm_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.08 (+5.07%)0.06 (+16.15%)0.06 (+1.68%)0.05 (+38.91%)0.01 (-42.74%)159.70 (-28.00%)134.52 (-18.84%)138.90 (-1.63%)108.90 (-4.81%)18.84 (-63.53%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.05 (n/a)0.06 (n/a)0.04 (n/a)0.02 (n/a)221.80 (n/a)165.74 (n/a)141.20 (n/a)114.40 (n/a)51.65 (n/a)

rms_norm_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-8.58%)0.05 (-4.03%)0.05 (-6.57%)0.05 (+24.03%)0.01 (-46.16%)176.70 (-19.35%)156.46 (+1.02%)157.90 (+7.05%)129.70 (+9.36%)16.93 (-55.52%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)219.10 (n/a)154.88 (n/a)147.50 (n/a)118.60 (n/a)38.07 (n/a)

rms_norm_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.08 (-4.60%)0.06 (+2.86%)0.06 (+7.31%)0.05 (-4.67%)0.01 (+4.09%)175.20 (+4.91%)137.86 (-2.19%)130.90 (-6.83%)108.20 (+4.84%)28.23 (+18.59%)
87e12ba — 2025-12-16 21:55:160.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)167.00 (n/a)140.94 (n/a)140.50 (n/a)103.20 (n/a)23.81 (n/a)

rms_norm_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+16.90%)0.06 (-1.43%)0.07 (+6.91%)0.04 (-35.74%)0.02 (+321.32%)231.60 (+55.65%)148.98 (+8.48%)125.80 (-6.47%)110.80 (-14.51%)49.24 (+476.15%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.06 (n/a)0.06 (n/a)0.06 (n/a)0.00 (n/a)148.80 (n/a)137.34 (n/a)134.50 (n/a)129.60 (n/a)8.55 (n/a)

rms_norm_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.08 (+7.82%)0.06 (+4.07%)0.06 (+2.48%)0.04 (-3.84%)0.01 (+21.06%)193.10 (+3.98%)138.78 (-2.60%)130.10 (-2.40%)100.20 (-7.22%)34.97 (+17.46%)
87e12ba — 2025-12-16 21:55:160.08 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)185.70 (n/a)142.48 (n/a)133.30 (n/a)108.00 (n/a)29.77 (n/a)

rms_norm_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+0.54%)0.06 (+19.36%)0.06 (+31.43%)0.05 (+52.09%)0.01 (-29.44%)163.00 (-34.25%)137.56 (-19.63%)131.90 (-23.89%)116.00 (-0.60%)22.73 (-54.01%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.05 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)247.90 (n/a)171.16 (n/a)173.30 (n/a)116.70 (n/a)49.41 (n/a)

rms_norm_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+36.89%)0.06 (+31.08%)0.06 (+30.25%)0.04 (+22.68%)0.01 (+69.39%)189.10 (-18.46%)149.78 (-22.73%)142.90 (-23.21%)117.90 (-26.95%)29.87 (+0.54%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.04 (n/a)0.04 (n/a)0.04 (n/a)0.01 (n/a)231.90 (n/a)193.84 (n/a)186.10 (n/a)161.40 (n/a)29.71 (n/a)

rms_norm_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+29.92%)0.06 (+25.61%)0.06 (+23.49%)0.05 (+27.70%)0.01 (+29.21%)177.20 (-21.70%)145.74 (-20.38%)145.40 (-19.00%)121.20 (-23.00%)20.39 (-22.78%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)226.30 (n/a)183.04 (n/a)179.50 (n/a)157.40 (n/a)26.40 (n/a)

rope_1_cols_2_channels_4096_tile_4096_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.21 (-9.02%)0.18 (-4.45%)0.18 (-9.56%)0.16 (+18.25%)0.02 (-36.81%)158.20 (-15.45%)135.62 (+2.15%)135.30 (+10.54%)115.70 (+9.98%)17.83 (-44.00%)
87e12ba — 2025-12-16 21:55:160.23 (n/a)0.19 (n/a)0.20 (n/a)0.13 (n/a)0.04 (n/a)187.10 (n/a)132.76 (n/a)122.40 (n/a)105.20 (n/a)31.84 (n/a)

rope_2_cols_2_channels_4096_tile_2048_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.20 (-2.80%)0.18 (-2.10%)0.18 (+5.19%)0.15 (-4.90%)0.02 (-5.65%)161.90 (+5.13%)139.78 (+2.10%)137.80 (-4.90%)120.50 (+2.90%)16.65 (+3.12%)
87e12ba — 2025-12-16 21:55:160.21 (n/a)0.18 (n/a)0.17 (n/a)0.16 (n/a)0.02 (n/a)154.00 (n/a)136.90 (n/a)144.90 (n/a)117.10 (n/a)16.15 (n/a)

rope_4_cols_2_channels_4096_tile_1024_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.22 (-0.71%)0.18 (-9.16%)0.17 (-12.69%)0.16 (-9.82%)0.02 (+23.34%)157.10 (+10.87%)140.60 (+10.66%)144.80 (+14.56%)112.20 (+0.72%)16.94 (+32.89%)
87e12ba — 2025-12-16 21:55:160.22 (n/a)0.20 (n/a)0.19 (n/a)0.17 (n/a)0.02 (n/a)141.70 (n/a)127.06 (n/a)126.40 (n/a)111.40 (n/a)12.74 (n/a)

rope_8_cols_2_channels_4096_tile_512_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.15 (-31.38%)0.14 (-30.78%)0.14 (-30.76%)0.11 (-31.91%)0.01 (-26.61%)221.00 (+46.84%)183.96 (+44.65%)177.70 (+44.35%)164.90 (+45.67%)22.29 (+56.36%)
87e12ba — 2025-12-16 21:55:160.22 (n/a)0.20 (n/a)0.20 (n/a)0.16 (n/a)0.02 (n/a)150.50 (n/a)127.18 (n/a)123.10 (n/a)113.20 (n/a)14.26 (n/a)

sigmoid_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-8.74%)0.06 (-8.17%)0.06 (-8.92%)0.05 (-9.20%)0.01 (-5.82%)162.10 (+10.12%)138.32 (+9.00%)133.90 (+9.84%)118.00 (+9.56%)17.61 (+13.49%)
87e12ba — 2025-12-16 21:55:160.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)147.20 (n/a)126.90 (n/a)121.90 (n/a)107.70 (n/a)15.52 (n/a)

sigmoid_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-1.84%)0.05 (-3.19%)0.05 (-1.20%)0.04 (-3.50%)0.01 (+11.07%)200.10 (+3.63%)161.36 (+3.92%)159.80 (+1.20%)129.30 (+1.81%)29.71 (+16.54%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)193.10 (n/a)155.28 (n/a)157.90 (n/a)127.00 (n/a)25.49 (n/a)

sigmoid_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-15.45%)0.05 (-11.19%)0.05 (-14.10%)0.05 (-1.40%)0.00 (-46.18%)167.70 (+1.45%)156.20 (+11.73%)161.00 (+16.41%)140.10 (+18.33%)11.14 (-35.92%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)165.30 (n/a)139.80 (n/a)138.30 (n/a)118.40 (n/a)17.39 (n/a)

sigmoid_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-12.44%)0.05 (-13.50%)0.05 (-20.04%)0.04 (-13.13%)0.01 (+0.10%)194.60 (+15.08%)164.30 (+16.54%)177.70 (+25.05%)121.70 (+14.27%)30.39 (+35.34%)
87e12ba — 2025-12-16 21:55:160.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)169.10 (n/a)140.98 (n/a)142.10 (n/a)106.50 (n/a)22.45 (n/a)

silu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.10 (+39.28%)0.07 (+13.05%)0.06 (+7.20%)0.05 (-1.40%)0.02 (+150.15%)162.40 (+1.44%)127.86 (-8.23%)126.10 (-6.66%)84.40 (-28.23%)28.89 (+75.59%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)160.10 (n/a)139.32 (n/a)135.10 (n/a)117.60 (n/a)16.45 (n/a)

silu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-5.74%)0.05 (-14.06%)0.05 (-14.16%)0.04 (-33.74%)0.01 (+130.15%)215.00 (+50.98%)158.00 (+19.72%)154.10 (+16.48%)125.60 (+6.08%)34.01 (+285.96%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.06 (n/a)0.00 (n/a)142.40 (n/a)131.98 (n/a)132.30 (n/a)118.40 (n/a)8.81 (n/a)

silu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+1.75%)0.06 (-8.90%)0.07 (-4.47%)0.04 (-31.60%)0.01 (+116.16%)204.70 (+46.21%)142.84 (+14.47%)123.10 (+4.68%)111.00 (-1.68%)38.35 (+211.87%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)140.00 (n/a)124.78 (n/a)117.60 (n/a)112.90 (n/a)12.30 (n/a)

silu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+2.56%)0.06 (+5.84%)0.07 (+10.65%)0.05 (+3.92%)0.01 (-8.87%)157.80 (-3.78%)130.56 (-5.79%)125.80 (-9.63%)116.20 (-2.52%)16.05 (-11.53%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)164.00 (n/a)138.58 (n/a)139.20 (n/a)119.20 (n/a)18.14 (n/a)

softmax_1_cols_2_channels_4096_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.13 (-18.44%)0.12 (+6.96%)0.12 (+11.03%)0.12 (+48.40%)0.00 (-89.40%)135.20 (-32.60%)131.84 (-10.45%)131.30 (-9.95%)128.20 (+22.56%)3.05 (-91.22%)
87e12ba — 2025-12-16 21:55:160.16 (n/a)0.12 (n/a)0.11 (n/a)0.08 (n/a)0.03 (n/a)200.60 (n/a)147.22 (n/a)145.80 (n/a)104.60 (n/a)34.71 (n/a)

softmax_2_cols_2_channels_4096_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.16 (+5.91%)0.11 (-12.30%)0.11 (-15.57%)0.07 (-37.33%)0.03 (+85.35%)233.70 (+59.52%)152.58 (+20.31%)143.70 (+18.47%)105.20 (-5.57%)48.11 (+192.82%)
87e12ba — 2025-12-16 21:55:160.15 (n/a)0.13 (n/a)0.14 (n/a)0.11 (n/a)0.02 (n/a)146.50 (n/a)126.82 (n/a)121.30 (n/a)111.40 (n/a)16.43 (n/a)

softmax_2_cols_2_channels_4096_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.14 (-13.29%)0.12 (-4.56%)0.12 (+6.03%)0.10 (+0.86%)0.01 (-49.64%)159.70 (-0.87%)136.06 (+2.62%)133.70 (-5.71%)118.80 (+15.34%)14.82 (-40.52%)
87e12ba — 2025-12-16 21:55:160.16 (n/a)0.13 (n/a)0.12 (n/a)0.10 (n/a)0.03 (n/a)161.10 (n/a)132.58 (n/a)141.80 (n/a)103.00 (n/a)24.92 (n/a)

swiglu

No metrics available.

swiglu_decode_1x2048x2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.00 (-15.46%)0.00 (-27.20%)0.00 (-28.17%)0.00 (-33.90%)0.00 (+1834.83%)5187.20 (+51.28%)4720.84 (+38.38%)4750.90 (+39.22%)4015.10 (+18.29%)390.23 (+3298.45%)
0a6c11c — 2025-12-04 00:39:100.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)3428.80 (n/a)3411.62 (n/a)3412.60 (n/a)3394.20 (n/a)11.48 (n/a)

swiglu_decode_1x2048x2048_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.00 (+33.10%)0.00 (+22.90%)0.00 (+39.14%)0.00 (+2.35%)0.00 (+253.24%)5485.70 (-2.30%)4205.48 (-14.94%)3431.60 (-28.13%)3386.10 (-24.87%)1086.90 (+154.86%)
87e12ba — 2025-12-16 21:55:160.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)5614.80 (n/a)4944.40 (n/a)4774.90 (n/a)4506.90 (n/a)426.47 (n/a)

tanh_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-9.49%)0.06 (-12.66%)0.06 (-11.16%)0.05 (-15.97%)0.00 (+64.88%)163.40 (+19.01%)146.90 (+14.89%)141.50 (+12.57%)137.00 (+10.48%)11.78 (+114.48%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.07 (n/a)0.06 (n/a)0.00 (n/a)137.30 (n/a)127.86 (n/a)125.70 (n/a)124.00 (n/a)5.49 (n/a)

tanh_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-8.90%)0.05 (-11.18%)0.06 (-9.66%)0.05 (-9.07%)0.01 (+4.31%)172.10 (+9.97%)150.82 (+12.87%)143.20 (+10.75%)131.70 (+9.75%)17.26 (+25.23%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)156.50 (n/a)133.62 (n/a)129.30 (n/a)120.00 (n/a)13.78 (n/a)

tanh_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.06 (-14.53%)0.05 (-23.31%)0.05 (-21.73%)0.03 (-41.29%)0.01 (+44.33%)268.10 (+70.33%)174.32 (+37.35%)151.70 (+27.80%)129.30 (+17.01%)56.53 (+191.57%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)157.40 (n/a)126.92 (n/a)118.70 (n/a)110.50 (n/a)19.39 (n/a)

tanh_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (-8.12%)0.05 (-14.67%)0.05 (-26.11%)0.04 (-16.15%)0.01 (-17.08%)208.10 (+19.26%)161.32 (+16.43%)168.00 (+35.37%)117.80 (+8.87%)33.98 (+3.21%)
87e12ba — 2025-12-16 21:55:160.08 (n/a)0.06 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)174.50 (n/a)138.56 (n/a)124.10 (n/a)108.20 (n/a)32.92 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:493.93 (-7.02%)3.68 (+2.48%)3.66 (+1.82%)3.45 (+15.62%)0.15 (-63.64%)151.80 (-13.50%)142.70 (-3.59%)143.40 (-1.78%)133.50 (+7.49%)5.90 (-66.22%)
0a6c11c — 2025-12-04 00:39:104.22 (n/a)3.59 (n/a)3.59 (n/a)2.99 (n/a)0.42 (n/a)175.50 (n/a)148.02 (n/a)146.00 (n/a)124.20 (n/a)17.46 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:334.07 (-2.05%)3.80 (+0.45%)3.81 (-3.10%)3.41 (+7.03%)0.25 (-35.23%)153.80 (-6.62%)138.32 (-0.99%)137.60 (+3.23%)128.70 (+2.06%)9.47 (-38.30%)
87e12ba — 2025-12-16 21:55:164.16 (n/a)3.79 (n/a)3.93 (n/a)3.18 (n/a)0.38 (n/a)164.70 (n/a)139.70 (n/a)133.30 (n/a)126.10 (n/a)15.35 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:495.11 (+15.57%)3.85 (-3.32%)3.40 (-18.64%)3.29 (+7.09%)0.71 (+49.14%)159.60 (-6.61%)140.18 (+4.80%)154.20 (+22.97%)102.60 (-13.49%)22.70 (+19.77%)
0a6c11c — 2025-12-04 00:39:104.42 (n/a)3.99 (n/a)4.18 (n/a)3.07 (n/a)0.48 (n/a)170.90 (n/a)133.76 (n/a)125.40 (n/a)118.60 (n/a)18.95 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:334.61 (-2.10%)3.65 (-9.33%)3.52 (-10.89%)2.82 (-11.30%)0.64 (+5.87%)185.80 (+12.74%)147.38 (+10.90%)149.00 (+12.28%)113.70 (+2.16%)25.89 (+22.29%)
87e12ba — 2025-12-16 21:55:164.71 (n/a)4.02 (n/a)3.95 (n/a)3.18 (n/a)0.61 (n/a)164.80 (n/a)132.90 (n/a)132.70 (n/a)111.30 (n/a)21.17 (n/a)

weighted_rms_norm_1_cols_2_channels_2048_weights_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+6.97%)0.06 (+14.87%)0.06 (+24.88%)0.05 (+12.52%)0.01 (-23.10%)155.80 (-11.12%)133.08 (-13.58%)130.20 (-19.93%)122.50 (-6.49%)13.59 (-34.56%)
87e12ba — 2025-12-16 21:55:160.06 (n/a)0.05 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)175.30 (n/a)154.00 (n/a)162.60 (n/a)131.00 (n/a)20.77 (n/a)

weighted_rms_norm_2_cols_2_channels_2048_weights_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+4.04%)0.06 (+3.81%)0.06 (+4.82%)0.04 (+7.73%)0.01 (+1.22%)196.40 (-7.18%)148.82 (-4.01%)146.40 (-4.56%)110.20 (-3.92%)32.11 (-10.70%)
87e12ba — 2025-12-16 21:55:160.07 (n/a)0.06 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)211.60 (n/a)155.04 (n/a)153.40 (n/a)114.70 (n/a)35.95 (n/a)

weighted_rms_norm_4_cols_2_channels_2048_weights_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+22.96%)0.06 (+37.07%)0.06 (+50.53%)0.05 (+54.39%)0.01 (-38.22%)149.60 (-35.24%)134.96 (-28.69%)134.60 (-33.56%)122.50 (-18.66%)11.73 (-66.59%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.04 (n/a)0.04 (n/a)0.04 (n/a)0.01 (n/a)231.00 (n/a)189.26 (n/a)202.60 (n/a)150.60 (n/a)35.10 (n/a)

weighted_rms_norm_8_cols_2_channels_2048_weights_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
68acccf — 2025-12-16 23:07:330.07 (+35.96%)0.05 (+5.35%)0.05 (-5.65%)0.03 (-12.05%)0.01 (+140.45%)237.80 (+13.73%)170.32 (-0.54%)172.90 (+6.01%)111.40 (-26.42%)45.81 (+97.01%)
87e12ba — 2025-12-16 21:55:160.05 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)209.10 (n/a)171.24 (n/a)163.10 (n/a)151.40 (n/a)23.25 (n/a)

@github-actions
Copy link
Contributor

📊 Test Results for Test Example Applications

b7882b4 (2025_12_18_22_39_37)

IRONCLAD

Tested on 2025_12_18_22_39_37 at commit b7882b4.

Test Checks TTFT (mean)TPS (mean)
llama_3.2_1b_prompt_13_tokens_1 ✅ 1/1 0.61 0.00
llama_3.2_1b_prompt_13_tokens_40 ✅ 1/1 0.62 4.77
llama_3.2_1b_prompt_2048_tokens_1 ✅ 1/1 4.38 0.00
llama_3.2_1b_prompt_2048_tokens_40 ✅ 1/1 4.41 4.76
📈 Trends (vs main branch) for Test Example Applications

b7882b4 (2025_12_18_22_39_37)

IRONCLAD Trends

llama_3.2_1b

Commit/Date Num Tokens (max)Num Tokens (mean)Num Tokens (median)Num Tokens (min)Num Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)Total (max)Total (mean)Total (median)Total (min)Total (stddev)
130b6ea — 2025-12-05 21:33:1240.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)0.00 (n/a)4.71 (-0.42%)4.64 (-0.09%)4.64 (+0.65%)4.55 (-0.22%)0.05 (-17.66%)4.41 (-0.34%)4.39 (-0.19%)4.38 (-0.33%)4.37 (-0.15%)0.01 (-25.90%)12.96 (-0.00%)12.80 (+0.07%)12.80 (-0.23%)12.67 (+0.44%)0.09 (-21.12%)
0a6c11c — 2025-12-03 23:35:1540.00 (n/a)40.00 (n/a)40.00 (n/a)40.00 (n/a)0.00 (n/a)4.73 (n/a)4.64 (n/a)4.61 (n/a)4.56 (n/a)0.06 (n/a)4.42 (n/a)4.40 (n/a)4.40 (n/a)4.37 (n/a)0.02 (n/a)12.96 (n/a)12.79 (n/a)12.83 (n/a)12.62 (n/a)0.12 (n/a)

llama_3.2_1b_prompt_13_tokens_1

Commit/Date Num_Tokens (max)Num_Tokens (mean)Num_Tokens (median)Num_Tokens (min)Num_Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)
b7882b4 — 2025-12-18 22:34:311.00 (+0.00%)1.00 (+0.00%)1.00 (+0.00%)1.00 (+0.00%)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.61 (-2.84%)0.61 (-2.84%)0.61 (-2.84%)0.61 (-2.84%)0.00 (n/a)
87e12ba — 2025-12-16 22:04:541.00 (n/a)1.00 (n/a)1.00 (n/a)1.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.63 (n/a)0.63 (n/a)0.63 (n/a)0.63 (n/a)0.00 (n/a)

llama_3.2_1b_prompt_13_tokens_40

Commit/Date Num_Tokens (max)Num_Tokens (mean)Num_Tokens (median)Num_Tokens (min)Num_Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)
b7882b4 — 2025-12-18 22:34:3140.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)0.00 (n/a)4.77 (-2.45%)4.77 (-2.45%)4.77 (-2.45%)4.77 (-2.45%)0.00 (n/a)0.62 (-1.87%)0.62 (-1.87%)0.62 (-1.87%)0.62 (-1.87%)0.00 (n/a)
87e12ba — 2025-12-16 22:04:5440.00 (n/a)40.00 (n/a)40.00 (n/a)40.00 (n/a)0.00 (n/a)4.89 (n/a)4.89 (n/a)4.89 (n/a)4.89 (n/a)0.00 (n/a)0.64 (n/a)0.64 (n/a)0.64 (n/a)0.64 (n/a)0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_1

Commit/Date Num_Tokens (max)Num_Tokens (mean)Num_Tokens (median)Num_Tokens (min)Num_Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)
b7882b4 — 2025-12-18 22:34:311.00 (+0.00%)1.00 (+0.00%)1.00 (+0.00%)1.00 (+0.00%)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)4.38 (-0.35%)4.38 (-0.35%)4.38 (-0.35%)4.38 (-0.35%)0.00 (n/a)
87e12ba — 2025-12-16 22:04:541.00 (n/a)1.00 (n/a)1.00 (n/a)1.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)4.39 (n/a)4.39 (n/a)4.39 (n/a)4.39 (n/a)0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_40

Commit/Date Num_Tokens (max)Num_Tokens (mean)Num_Tokens (median)Num_Tokens (min)Num_Tokens (stddev)TPS (max)TPS (mean)TPS (median)TPS (min)TPS (stddev)TTFT (max)TTFT (mean)TTFT (median)TTFT (min)TTFT (stddev)
b7882b4 — 2025-12-18 22:34:3140.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)40.00 (+0.00%)0.00 (n/a)4.76 (+5.31%)4.76 (+5.31%)4.76 (+5.31%)4.76 (+5.31%)0.00 (n/a)4.41 (+1.05%)4.41 (+1.05%)4.41 (+1.05%)4.41 (+1.05%)0.00 (n/a)
87e12ba — 2025-12-16 22:04:5440.00 (n/a)40.00 (n/a)40.00 (n/a)40.00 (n/a)0.00 (n/a)4.52 (n/a)4.52 (n/a)4.52 (n/a)4.52 (n/a)0.00 (n/a)4.36 (n/a)4.36 (n/a)4.36 (n/a)4.36 (n/a)0.00 (n/a)

@github-actions
Copy link
Contributor

📊 Test Results for Small Benchmark/Test Suite

b7882b4 (2025_12_18_23_21_22)

IRONCLAD

Tested on 2025_12_18_23_21_22 at commit b7882b4.

Test Checks Latency (mean)Bandwidth (mean)GFLOP/s (mean)
axpy_1_cols_2_channels_2048_tile_2048_3.0_0 ✅ 5/5 123.52 0.10 n/a
axpy_2_cols_2_channels_2048_tile_1024_3.0_0 ✅ 5/5 136.84 0.10 n/a
axpy_4_cols_2_channels_2048_tile_512_3.0_0 ✅ 5/5 126.94 0.10 n/a
axpy_8_cols_2_channels_2048_tile_256_3.0_0 ✅ 5/5 152.00 0.08 n/a
dequant_1_cols_1_channels_2048_tile_2048_0 ✅ 5/5 120.38 0.04 n/a
dequant_1_cols_2_channels_2048_tile_1024_0 ✅ 5/5 127.80 0.04 n/a
dequant_2_cols_1_channels_2048_tile_1024_0 ✅ 5/5 124.44 0.04 n/a
dequant_2_cols_2_channels_2048_tile_512_0 ✅ 5/5 131.32 0.04 n/a
dequant_4_cols_1_channels_2048_tile_512_0 ✅ 5/5 118.60 0.05 n/a
dequant_4_cols_2_channels_2048_tile_256_0 ✅ 5/5 136.92 0.04 n/a
dequant_8_cols_1_channels_2048_tile_256_0 ✅ 5/5 125.30 0.04 n/a
dequant_8_cols_2_channels_2048_tile_128_0 ✅ 5/5 154.94 0.03 n/a
eltwise_add_1_cols_2_channels_2048_tile_2048 ✅ 5/5 146.50 0.09 n/a
eltwise_add_2_cols_2_channels_2048_tile_1024 ✅ 5/5 133.98 0.09 n/a
eltwise_add_4_cols_2_channels_2048_tile_512 ✅ 5/5 144.92 0.09 n/a
eltwise_add_8_cols_2_channels_2048_tile_256 ✅ 5/5 158.92 0.08 n/a
eltwise_mul_1_cols_2_channels_2048_tile_2048 ✅ 5/5 123.56 0.10 n/a
eltwise_mul_2_cols_2_channels_2048_tile_1024 ✅ 5/5 137.00 0.09 n/a
eltwise_mul_4_cols_2_channels_2048_tile_512 ✅ 5/5 118.22 0.11 n/a
eltwise_mul_8_cols_2_channels_2048_tile_256 ✅ 5/5 159.36 0.08 n/a
gelu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 143.62 0.06 n/a
gelu_1_cols_2_channels_2048_tile_1024 ✅ 5/5 153.48 0.05 n/a
gelu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 159.38 0.05 n/a
gelu_2_cols_2_channels_2048_tile_512 ✅ 5/5 160.56 0.05 n/a
gelu_4_cols_1_channels_2048_tile_512 ✅ 5/5 130.60 0.06 n/a
gelu_4_cols_2_channels_2048_tile_256 ✅ 5/5 148.34 0.06 n/a
gelu_8_cols_1_channels_2048_tile_256 ✅ 5/5 151.44 0.05 n/a
gelu_8_cols_2_channels_2048_tile_128 ✅ 5/5 184.42 0.05 n/a
gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0 ✅ 5/5 24524.60 1.03 n/a
gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0 ✅ 5/5 25903.50 0.97 n/a
gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0 ✅ 5/5 27627.92 0.91 n/a
layer_norm_1_cols_1_channels_2048_tile_2048 ✅ 5/5 145.30 0.06 n/a
layer_norm_1_cols_2_channels_2048_tile_1024 ✅ 5/5 140.68 0.06 n/a
layer_norm_2_cols_1_channels_2048_tile_1024 ✅ 5/5 142.30 0.06 n/a
layer_norm_2_cols_2_channels_2048_tile_512 ✅ 5/5 157.54 0.06 n/a
layer_norm_4_cols_1_channels_2048_tile_512 ✅ 5/5 184.64 0.05 n/a
layer_norm_4_cols_2_channels_2048_tile_256 ✅ 5/5 157.18 0.05 n/a
layer_norm_8_cols_1_channels_2048_tile_256 ✅ 5/5 175.32 0.05 n/a
layer_norm_8_cols_2_channels_2048_tile_128 ✅ 5/5 184.40 0.04 n/a
matrix_vector_mul_128x128_32_1col0 ✅ 5/5 n/a 0.25 n/a
matrix_vector_mul_2048x8192_1_1col0 ✅ 5/5 n/a 12.45 n/a
matrix_vector_mul_2048x8192_1_2col0 ✅ 5/5 n/a 24.46 n/a
matrix_vector_mul_2048x8192_1_4col0 ✅ 5/5 n/a 39.57 n/a
matrix_vector_mul_2048x8192_1_8col0 ✅ 5/5 n/a 41.60 n/a
matrix_vector_mul_8192x2048_4_1col0 ✅ 5/5 n/a 12.03 n/a
matrix_vector_mul_8192x2048_4_2col0 ✅ 5/5 n/a 24.33 n/a
matrix_vector_mul_8192x2048_4_4col0 ✅ 5/5 n/a 39.82 n/a
matrix_vector_mul_8192x2048_4_8col0 ✅ 5/5 n/a 44.03 n/a
mem_copy_16_cores_2_chans_2048_tile_128_False0 ✅ 5/5 165.76 0.05 n/a
mem_copy_1_cores_1_chans_2048_tile_2048_False0 ✅ 5/5 136.88 0.06 n/a
mem_copy_2_cores_1_chans_2048_tile_1024_False0 ✅ 5/5 163.34 0.06 n/a
mem_copy_2_cores_2_chans_2048_tile_1024_False0 ✅ 5/5 142.36 0.06 n/a
mem_copy_4_cores_1_chans_2048_tile_512_False0 ✅ 5/5 152.50 0.05 n/a
mem_copy_4_cores_2_chans_2048_tile_512_False0 ✅ 5/5 151.50 0.06 n/a
mem_copy_8_cores_1_chans_2048_tile_256_False0 ✅ 5/5 149.40 0.06 n/a
mem_copy_8_cores_2_chans_2048_tile_256_False0 ✅ 5/5 178.42 0.05 n/a
mha0 ✅ 5/5 47196.86 0.18 n/a
relu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 155.58 0.05 n/a
relu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 139.72 0.06 n/a
relu_4_cols_1_channels_2048_tile_512 ✅ 5/5 144.32 0.06 n/a
relu_8_cols_1_channels_2048_tile_256 ✅ 5/5 159.58 0.05 n/a
rms_norm_1_cols_1_channels_2048_tile_2048 ✅ 5/5 137.74 0.06 n/a
rms_norm_1_cols_2_channels_2048_tile_1024 ✅ 5/5 150.34 0.06 n/a
rms_norm_2_cols_1_channels_2048_tile_1024 ✅ 5/5 149.54 0.06 n/a
rms_norm_2_cols_2_channels_2048_tile_512 ✅ 5/5 151.16 0.06 n/a
rms_norm_4_cols_1_channels_2048_tile_512 ✅ 5/5 157.42 0.05 n/a
rms_norm_4_cols_2_channels_2048_tile_256 ✅ 5/5 137.42 0.06 n/a
rms_norm_8_cols_1_channels_2048_tile_256 ✅ 5/5 184.32 0.05 n/a
rms_norm_8_cols_2_channels_2048_tile_128 ✅ 5/5 163.00 0.05 n/a
rope_1_cols_2_channels_4096_tile_4096_0 ✅ 5/5 122.54 0.20 n/a
rope_2_cols_2_channels_4096_tile_2048_0 ✅ 5/5 133.10 0.19 n/a
rope_4_cols_2_channels_4096_tile_1024_0 ✅ 5/5 118.94 0.21 n/a
rope_8_cols_2_channels_4096_tile_512_0 ✅ 5/5 142.38 0.18 n/a
sigmoid_1_cols_1_channels_2048_tile_2048 ✅ 5/5 138.70 0.06 n/a
sigmoid_2_cols_1_channels_2048_tile_1024 ✅ 5/5 133.90 0.06 n/a
sigmoid_4_cols_1_channels_2048_tile_512 ✅ 5/5 144.22 0.06 n/a
sigmoid_8_cols_1_channels_2048_tile_256 ✅ 5/5 139.48 0.06 n/a
silu_1_cols_1_channels_2048_tile_2048 ✅ 5/5 140.46 0.06 n/a
silu_2_cols_1_channels_2048_tile_1024 ✅ 5/5 155.06 0.05 n/a
silu_4_cols_1_channels_2048_tile_512 ✅ 5/5 162.18 0.05 n/a
silu_8_cols_1_channels_2048_tile_256 ✅ 5/5 163.56 0.05 n/a
softmax_1_cols_2_channels_4096_tile_2048 ✅ 5/5 135.68 0.13 n/a
softmax_2_cols_2_channels_4096_tile_1024 ✅ 5/5 128.52 0.13 n/a
softmax_2_cols_2_channels_4096_tile_512 ✅ 5/5 133.34 0.14 n/a
swiglu_decode_1x2048x2048_0 ✅ 5/5 4533.84 0.00 n/a
tanh_1_cols_1_channels_2048_tile_2048 ✅ 5/5 133.72 0.06 n/a
tanh_2_cols_1_channels_2048_tile_1024 ✅ 5/5 161.68 0.05 n/a
tanh_4_cols_1_channels_2048_tile_512 ✅ 5/5 148.56 0.06 n/a
tanh_8_cols_1_channels_2048_tile_256 ✅ 5/5 208.80 0.04 n/a
transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0 ✅ 5/5 137.76 3.81 n/a
transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0 ✅ 5/5 119.24 4.43 n/a
weighted_rms_norm_1_cols_2_channels_2048_weights_2048 ✅ 5/5 150.38 0.06 n/a
weighted_rms_norm_2_cols_2_channels_2048_weights_1024 ✅ 5/5 122.40 0.07 n/a
weighted_rms_norm_4_cols_2_channels_2048_weights_512 ✅ 5/5 121.90 0.07 n/a
weighted_rms_norm_8_cols_2_channels_2048_weights_256 ✅ 5/5 153.90 0.05 n/a
📈 Trends (vs main branch) for Small Benchmark/Test Suite

b7882b4 (2025_12_18_23_21_22)

IRONCLAD Trends

axpy_1_cols_2_channels_2048_tile_2048_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.13 (-1.96%)0.11 (+2.52%)0.10 (+1.86%)0.08 (-1.91%)0.02 (-1.72%)162.40 (+1.95%)120.96 (-2.36%)117.00 (-1.85%)94.40 (+2.05%)23.11 (+4.09%)
0a6c11c — 2025-12-04 00:39:100.13 (n/a)0.10 (n/a)0.10 (n/a)0.08 (n/a)0.02 (n/a)159.30 (n/a)123.88 (n/a)119.20 (n/a)92.50 (n/a)22.21 (n/a)

axpy_1_cols_2_channels_2048_tile_2048_3.0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.11 (+21.98%)0.10 (+22.94%)0.11 (+32.19%)0.09 (+12.49%)0.01 (+100.77%)141.30 (-11.08%)123.52 (-18.27%)115.90 (-24.35%)114.50 (-17.98%)11.85 (+44.53%)
ea7f586 — 2025-12-18 22:24:430.09 (n/a)0.08 (n/a)0.08 (n/a)0.08 (n/a)0.00 (n/a)158.90 (n/a)151.14 (n/a)153.20 (n/a)139.60 (n/a)8.20 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.11 (-26.77%)0.09 (-9.09%)0.09 (-10.81%)0.08 (+8.25%)0.01 (-55.42%)152.00 (-7.60%)135.46 (+6.28%)142.80 (+12.18%)115.00 (+36.58%)15.03 (-41.36%)
0a6c11c — 2025-12-04 00:39:100.15 (n/a)0.10 (n/a)0.10 (n/a)0.07 (n/a)0.02 (n/a)164.50 (n/a)127.46 (n/a)127.30 (n/a)84.20 (n/a)25.64 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.13 (+13.91%)0.10 (+3.00%)0.11 (+18.44%)0.06 (-24.21%)0.03 (+91.21%)204.10 (+31.93%)136.84 (+2.70%)114.10 (-15.54%)97.10 (-12.21%)44.32 (+123.81%)
ea7f586 — 2025-12-18 22:24:430.11 (n/a)0.09 (n/a)0.09 (n/a)0.08 (n/a)0.01 (n/a)154.70 (n/a)133.24 (n/a)135.10 (n/a)110.60 (n/a)19.80 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.12 (-10.21%)0.10 (-8.41%)0.10 (-6.20%)0.07 (+3.52%)0.02 (-23.48%)166.70 (-3.36%)128.78 (+7.33%)121.80 (+6.65%)103.90 (+11.36%)22.95 (-18.82%)
0a6c11c — 2025-12-04 00:39:100.13 (n/a)0.11 (n/a)0.11 (n/a)0.07 (n/a)0.02 (n/a)172.50 (n/a)119.98 (n/a)114.20 (n/a)93.30 (n/a)28.27 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.11 (-7.42%)0.10 (+7.60%)0.10 (+4.02%)0.08 (+15.22%)0.01 (-44.51%)147.10 (-13.16%)126.94 (-9.32%)129.30 (-3.87%)109.20 (+8.01%)14.26 (-48.71%)
ea7f586 — 2025-12-18 22:24:430.12 (n/a)0.09 (n/a)0.09 (n/a)0.07 (n/a)0.02 (n/a)169.40 (n/a)139.98 (n/a)134.50 (n/a)101.10 (n/a)27.81 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.10 (-16.19%)0.08 (-10.40%)0.09 (-2.09%)0.07 (-17.38%)0.01 (-10.61%)184.40 (+21.00%)153.18 (+11.99%)144.10 (+2.20%)127.10 (+19.34%)21.26 (+34.76%)
0a6c11c — 2025-12-04 00:39:100.12 (n/a)0.09 (n/a)0.09 (n/a)0.08 (n/a)0.01 (n/a)152.40 (n/a)136.78 (n/a)141.00 (n/a)106.50 (n/a)15.78 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.09 (-7.27%)0.08 (+5.05%)0.08 (+23.90%)0.07 (+4.06%)0.01 (-39.18%)181.50 (-3.92%)152.00 (-6.95%)148.70 (-19.32%)129.60 (+7.91%)20.28 (-38.54%)
ea7f586 — 2025-12-18 22:24:430.10 (n/a)0.08 (n/a)0.07 (n/a)0.07 (n/a)0.02 (n/a)188.90 (n/a)163.36 (n/a)184.30 (n/a)120.10 (n/a)33.00 (n/a)

dequant_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.05 (-18.83%)0.04 (-5.21%)0.04 (-2.69%)0.03 (+37.10%)0.00 (-53.98%)160.40 (-27.06%)141.12 (-0.91%)136.40 (+2.79%)114.30 (+23.17%)17.37 (-59.06%)
0a6c11c — 2025-12-04 00:39:100.06 (n/a)0.04 (n/a)0.04 (n/a)0.02 (n/a)0.01 (n/a)219.90 (n/a)142.42 (n/a)132.70 (n/a)92.80 (n/a)42.43 (n/a)

dequant_1_cols_1_channels_2048_tile_2048_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (-6.81%)0.04 (+0.06%)0.05 (-0.76%)0.04 (+5.45%)0.01 (-19.15%)142.70 (-5.12%)120.38 (-1.18%)108.60 (+0.74%)104.10 (+7.32%)20.23 (-19.78%)
ea7f586 — 2025-12-18 22:24:430.05 (n/a)0.04 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)150.40 (n/a)121.82 (n/a)107.80 (n/a)97.00 (n/a)25.21 (n/a)

dequant_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.06 (-13.28%)0.05 (-8.93%)0.06 (-13.59%)0.04 (+4.63%)0.01 (-39.10%)133.60 (-4.43%)106.08 (+5.85%)95.10 (+15.69%)87.90 (+15.35%)17.90 (-31.97%)
0a6c11c — 2025-12-04 00:39:100.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)139.80 (n/a)100.22 (n/a)82.20 (n/a)76.20 (n/a)26.31 (n/a)

dequant_1_cols_2_channels_2048_tile_1024_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (-0.21%)0.04 (+2.34%)0.04 (+5.61%)0.04 (+6.00%)0.00 (-22.10%)146.20 (-5.62%)127.80 (-3.05%)127.60 (-5.27%)111.20 (+0.27%)15.31 (-24.85%)
ea7f586 — 2025-12-18 22:24:430.05 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)154.90 (n/a)131.82 (n/a)134.70 (n/a)110.90 (n/a)20.38 (n/a)

dequant_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.07 (+5.66%)0.05 (+1.21%)0.05 (+0.61%)0.04 (-15.02%)0.01 (+58.42%)143.40 (+17.64%)101.02 (+1.49%)97.60 (-0.61%)78.90 (-5.40%)22.56 (+78.52%)
0a6c11c — 2025-12-04 00:39:100.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)121.90 (n/a)99.54 (n/a)98.20 (n/a)83.40 (n/a)12.64 (n/a)

dequant_2_cols_1_channels_2048_tile_1024_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (-26.78%)0.04 (-10.75%)0.04 (+3.55%)0.03 (-7.41%)0.01 (-51.50%)164.80 (+7.99%)124.44 (+7.55%)120.40 (-3.37%)104.10 (+36.61%)23.58 (-24.95%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.05 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)152.60 (n/a)115.70 (n/a)124.60 (n/a)76.20 (n/a)31.42 (n/a)

dequant_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.05 (+17.06%)0.04 (+8.98%)0.05 (+21.37%)0.03 (-11.65%)0.01 (+301.89%)159.70 (+13.18%)126.32 (-5.68%)109.00 (-17.67%)106.80 (-14.56%)22.48 (+280.15%)
0a6c11c — 2025-12-04 00:39:100.04 (n/a)0.04 (n/a)0.04 (n/a)0.04 (n/a)0.00 (n/a)141.10 (n/a)133.92 (n/a)132.40 (n/a)125.00 (n/a)5.91 (n/a)

dequant_2_cols_2_channels_2048_tile_512_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.04 (-14.98%)0.04 (-5.03%)0.04 (-12.87%)0.03 (+15.75%)0.00 (-55.57%)154.30 (-13.61%)131.32 (+1.17%)128.90 (+14.78%)117.40 (+17.64%)15.22 (-55.24%)
ea7f586 — 2025-12-18 22:24:430.05 (n/a)0.04 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)178.60 (n/a)129.80 (n/a)112.30 (n/a)99.80 (n/a)34.00 (n/a)

dequant_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.06 (+20.24%)0.05 (-3.19%)0.05 (-6.73%)0.03 (-20.69%)0.01 (+194.26%)156.10 (+26.09%)120.86 (+7.15%)114.90 (+7.18%)88.30 (-16.85%)24.40 (+214.82%)
0a6c11c — 2025-12-04 00:39:100.05 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.00 (n/a)123.80 (n/a)112.80 (n/a)107.20 (n/a)106.20 (n/a)7.75 (n/a)

dequant_4_cols_1_channels_2048_tile_512_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (-0.38%)0.05 (+15.75%)0.05 (+28.25%)0.03 (+6.35%)0.01 (-18.08%)153.40 (-6.00%)118.60 (-14.58%)115.70 (-21.98%)97.20 (+0.41%)21.19 (-18.86%)
ea7f586 — 2025-12-18 22:24:430.05 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)163.20 (n/a)138.84 (n/a)148.30 (n/a)96.80 (n/a)26.12 (n/a)

dequant_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.05 (-6.09%)0.04 (-8.04%)0.04 (-12.46%)0.03 (-0.41%)0.01 (-20.19%)168.10 (+0.42%)135.74 (+7.44%)132.20 (+14.26%)107.40 (+6.44%)22.91 (-12.67%)
0a6c11c — 2025-12-04 00:39:100.05 (n/a)0.04 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)167.40 (n/a)126.34 (n/a)115.70 (n/a)100.90 (n/a)26.23 (n/a)

dequant_4_cols_2_channels_2048_tile_256_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (+20.54%)0.04 (-0.42%)0.04 (-5.37%)0.03 (-20.10%)0.01 (+100.20%)200.30 (+25.11%)136.92 (+5.00%)126.70 (+5.67%)96.30 (-16.98%)39.39 (+113.15%)
ea7f586 — 2025-12-18 22:24:430.05 (n/a)0.04 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)160.10 (n/a)130.40 (n/a)119.90 (n/a)116.00 (n/a)18.48 (n/a)

dequant_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.04 (-13.65%)0.04 (-13.67%)0.04 (-18.56%)0.03 (-13.52%)0.00 (-23.88%)170.40 (+15.68%)140.12 (+15.33%)135.80 (+22.78%)119.60 (+15.78%)17.94 (+1.87%)
0a6c11c — 2025-12-04 00:39:100.05 (n/a)0.04 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)147.30 (n/a)121.50 (n/a)110.60 (n/a)103.30 (n/a)17.61 (n/a)

dequant_8_cols_1_channels_2048_tile_256_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (+29.34%)0.04 (+25.63%)0.04 (+14.45%)0.03 (+41.30%)0.01 (+13.77%)160.40 (-29.25%)125.30 (-21.54%)120.60 (-12.61%)94.00 (-22.70%)25.95 (-38.51%)
ea7f586 — 2025-12-18 22:24:430.04 (n/a)0.03 (n/a)0.04 (n/a)0.02 (n/a)0.01 (n/a)226.70 (n/a)159.70 (n/a)138.00 (n/a)121.60 (n/a)42.20 (n/a)

dequant_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.05 (+17.21%)0.03 (-1.70%)0.03 (-3.04%)0.03 (-16.20%)0.01 (+95.59%)198.20 (+19.33%)160.00 (+4.92%)166.20 (+3.17%)106.50 (-14.66%)29.86 (+87.19%)
0a6c11c — 2025-12-04 00:39:100.04 (n/a)0.03 (n/a)0.03 (n/a)0.03 (n/a)0.00 (n/a)166.10 (n/a)152.50 (n/a)161.10 (n/a)124.80 (n/a)15.95 (n/a)

dequant_8_cols_2_channels_2048_tile_128_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (+11.53%)0.03 (+7.22%)0.03 (-2.73%)0.03 (+20.71%)0.01 (+8.43%)178.60 (-17.16%)154.94 (-7.12%)169.10 (+2.80%)111.70 (-10.28%)27.77 (-19.71%)
ea7f586 — 2025-12-18 22:24:430.04 (n/a)0.03 (n/a)0.03 (n/a)0.02 (n/a)0.01 (n/a)215.60 (n/a)166.82 (n/a)164.50 (n/a)124.50 (n/a)34.59 (n/a)

eltwise_add_1_cols_2_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.10 (-22.15%)0.09 (-16.25%)0.09 (-2.28%)0.07 (-16.57%)0.01 (-45.60%)182.30 (+19.86%)146.50 (+17.16%)138.70 (+2.36%)122.60 (+28.38%)22.36 (-12.69%)
ea7f586 — 2025-12-18 22:24:430.13 (n/a)0.10 (n/a)0.09 (n/a)0.08 (n/a)0.02 (n/a)152.10 (n/a)125.04 (n/a)135.50 (n/a)95.50 (n/a)25.61 (n/a)

eltwise_add_2_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.12 (-4.07%)0.09 (-7.83%)0.09 (-1.19%)0.07 (-23.02%)0.02 (+46.50%)168.90 (+29.92%)133.98 (+10.24%)130.30 (+1.16%)106.30 (+4.22%)23.85 (+99.28%)
ea7f586 — 2025-12-18 22:24:430.12 (n/a)0.10 (n/a)0.10 (n/a)0.09 (n/a)0.01 (n/a)130.00 (n/a)121.54 (n/a)128.80 (n/a)102.00 (n/a)11.97 (n/a)

eltwise_add_4_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.11 (-0.00%)0.09 (+0.61%)0.09 (+3.57%)0.07 (-4.78%)0.02 (+3.35%)180.70 (+5.00%)144.92 (-0.29%)141.60 (-3.41%)107.80 (+0.00%)28.10 (+8.07%)
ea7f586 — 2025-12-18 22:24:430.11 (n/a)0.09 (n/a)0.08 (n/a)0.07 (n/a)0.02 (n/a)172.10 (n/a)145.34 (n/a)146.60 (n/a)107.80 (n/a)26.00 (n/a)

eltwise_add_8_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.10 (-5.99%)0.08 (-14.90%)0.07 (-15.60%)0.07 (-13.48%)0.01 (+13.03%)183.60 (+15.62%)158.92 (+18.61%)165.10 (+18.44%)120.00 (+6.38%)26.57 (+42.18%)
ea7f586 — 2025-12-18 22:24:430.11 (n/a)0.09 (n/a)0.09 (n/a)0.08 (n/a)0.01 (n/a)158.80 (n/a)133.98 (n/a)139.40 (n/a)112.80 (n/a)18.69 (n/a)

eltwise_mul_1_cols_2_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.11 (-2.14%)0.10 (+3.82%)0.10 (+2.45%)0.09 (+12.53%)0.01 (-34.12%)135.70 (-11.13%)123.56 (-4.42%)122.40 (-2.39%)110.50 (+2.13%)9.83 (-40.18%)
ea7f586 — 2025-12-18 22:24:430.11 (n/a)0.10 (n/a)0.10 (n/a)0.08 (n/a)0.01 (n/a)152.70 (n/a)129.28 (n/a)125.40 (n/a)108.20 (n/a)16.43 (n/a)

eltwise_mul_2_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.11 (-2.77%)0.09 (-0.85%)0.09 (-3.81%)0.07 (-8.71%)0.01 (+26.74%)167.10 (+9.57%)137.00 (+1.74%)142.20 (+3.95%)114.70 (+2.78%)21.71 (+41.12%)
ea7f586 — 2025-12-18 22:24:430.11 (n/a)0.09 (n/a)0.09 (n/a)0.08 (n/a)0.01 (n/a)152.50 (n/a)134.66 (n/a)136.80 (n/a)111.60 (n/a)15.38 (n/a)

eltwise_mul_4_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.14 (+20.62%)0.11 (+8.50%)0.10 (+3.74%)0.08 (-1.68%)0.02 (+49.33%)150.00 (+1.69%)118.22 (-6.60%)118.10 (-3.59%)89.40 (-17.07%)22.48 (+24.16%)
ea7f586 — 2025-12-18 22:24:430.11 (n/a)0.10 (n/a)0.10 (n/a)0.08 (n/a)0.01 (n/a)147.50 (n/a)126.58 (n/a)122.50 (n/a)107.80 (n/a)18.11 (n/a)

eltwise_mul_8_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.10 (-9.58%)0.08 (-6.31%)0.08 (-6.63%)0.06 (-15.15%)0.02 (-4.93%)211.60 (+17.82%)159.36 (+7.20%)148.60 (+7.06%)125.20 (+10.60%)34.10 (+21.54%)
ea7f586 — 2025-12-18 22:24:430.11 (n/a)0.09 (n/a)0.09 (n/a)0.07 (n/a)0.02 (n/a)179.60 (n/a)148.66 (n/a)138.80 (n/a)113.20 (n/a)28.06 (n/a)

gelu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-12.83%)0.06 (-2.58%)0.06 (+1.53%)0.05 (+1.40%)0.01 (-25.91%)168.90 (-1.40%)143.62 (+1.79%)138.40 (-1.49%)121.60 (+14.72%)20.32 (-13.22%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)171.30 (n/a)141.10 (n/a)140.50 (n/a)106.00 (n/a)23.42 (n/a)

gelu_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-12.59%)0.05 (-12.01%)0.05 (-11.19%)0.05 (-17.13%)0.01 (-3.94%)178.80 (+20.73%)153.48 (+13.96%)157.30 (+12.60%)127.90 (+14.40%)18.63 (+34.63%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.06 (n/a)0.01 (n/a)148.10 (n/a)134.68 (n/a)139.70 (n/a)111.80 (n/a)13.84 (n/a)

gelu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-7.53%)0.05 (-22.29%)0.05 (-25.73%)0.04 (-31.51%)0.01 (+47.51%)200.90 (+46.00%)159.38 (+31.35%)159.60 (+34.68%)117.80 (+8.17%)29.59 (+129.83%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)137.60 (n/a)121.34 (n/a)118.50 (n/a)108.90 (n/a)12.88 (n/a)

gelu_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-21.46%)0.05 (-15.65%)0.05 (-4.33%)0.04 (-9.57%)0.01 (-34.28%)200.50 (+10.59%)160.56 (+16.65%)151.10 (+4.57%)123.10 (+27.30%)29.80 (-5.52%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)181.30 (n/a)137.64 (n/a)144.50 (n/a)96.70 (n/a)31.54 (n/a)

gelu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-16.05%)0.06 (+1.53%)0.07 (+15.62%)0.05 (-5.43%)0.01 (-31.09%)174.10 (+5.71%)130.60 (-2.78%)123.80 (-13.49%)110.80 (+19.14%)24.98 (-6.63%)
ea7f586 — 2025-12-18 22:24:430.09 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)164.70 (n/a)134.34 (n/a)143.10 (n/a)93.00 (n/a)26.75 (n/a)

gelu_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-3.33%)0.06 (-1.84%)0.06 (+11.65%)0.04 (-3.83%)0.01 (-15.79%)205.60 (+3.94%)148.34 (+0.76%)142.50 (-10.43%)109.70 (+3.39%)36.50 (-4.26%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.05 (n/a)0.04 (n/a)0.02 (n/a)197.80 (n/a)147.22 (n/a)159.10 (n/a)106.10 (n/a)38.12 (n/a)

gelu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (+3.18%)0.05 (-0.44%)0.05 (-0.08%)0.05 (-0.25%)0.01 (-10.54%)174.10 (+0.29%)151.44 (-0.05%)156.00 (+0.06%)121.00 (-3.12%)19.44 (-15.63%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)173.60 (n/a)151.52 (n/a)155.90 (n/a)124.90 (n/a)23.04 (n/a)

gelu_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-5.34%)0.05 (-14.17%)0.05 (-5.63%)0.04 (-20.38%)0.01 (+16.68%)222.50 (+25.56%)184.42 (+18.28%)173.40 (+5.99%)135.20 (+5.71%)36.39 (+57.57%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.05 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)177.20 (n/a)155.92 (n/a)163.60 (n/a)127.90 (n/a)23.10 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.72 (-0.11%)3.55 (-2.75%)3.56 (-2.75%)3.36 (-5.33%)0.12 (+76.10%)2725.23 (-0.11%)2602.52 (-2.75%)2612.11 (-2.75%)2459.89 (-5.33%)86.27 (+76.10%)6984.00 (+5.63%)6608.60 (+2.91%)6577.00 (+2.83%)6304.00 (+0.11%)221.87 (+87.15%)
9e98049 — 2025-11-25 20:50:553.72 (n/a)3.65 (n/a)3.66 (n/a)3.54 (n/a)0.07 (n/a)2728.26 (n/a)2676.06 (n/a)2686.03 (n/a)2598.29 (n/a)48.99 (n/a)6612.00 (n/a)6422.00 (n/a)6396.00 (n/a)6297.00 (n/a)118.55 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.40 (-3.14%)3.38 (-0.82%)3.38 (-0.82%)3.35 (+4.93%)0.02 (-84.51%)2493.81 (-3.14%)2475.27 (-0.82%)2479.06 (-0.82%)2453.92 (+4.93%)13.33 (-84.51%)7001.00 (-4.70%)6940.80 (+0.71%)6930.00 (+0.83%)6889.00 (+3.24%)37.43 (-84.76%)
9e98049 — 2025-11-25 20:50:553.51 (n/a)3.40 (n/a)3.41 (n/a)3.19 (n/a)0.12 (n/a)2574.53 (n/a)2495.72 (n/a)2499.62 (n/a)2338.67 (n/a)86.02 (n/a)7346.00 (n/a)6892.20 (n/a)6873.00 (n/a)6673.00 (n/a)245.57 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
3a053d3 — 2025-11-26 01:12:163.02 (+0.52%)2.87 (-2.05%)2.86 (-2.24%)2.70 (-6.16%)0.11 (+131.66%)2216.19 (+0.52%)2104.44 (-2.05%)2094.08 (-2.24%)1980.84 (-6.16%)80.25 (+131.67%)8673.00 (+6.56%)8175.60 (+2.21%)8204.00 (+2.29%)7752.00 (-0.51%)314.03 (+145.36%)
9e98049 — 2025-11-25 20:50:553.01 (n/a)2.93 (n/a)2.92 (n/a)2.88 (n/a)0.05 (n/a)2204.81 (n/a)2148.41 (n/a)2142.13 (n/a)2110.81 (n/a)34.64 (n/a)8139.00 (n/a)7998.60 (n/a)8020.00 (n/a)7792.00 (n/a)127.99 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:491.04 (+0.94%)1.03 (+0.23%)1.03 (+0.09%)1.02 (+0.07%)0.00 (+197.51%)24576.00 (-0.07%)24485.82 (-0.23%)24530.40 (-0.09%)24266.40 (-0.94%)111.26 (+194.35%)
0a6c11c — 2025-12-04 00:39:101.03 (n/a)1.03 (n/a)1.03 (n/a)1.02 (n/a)0.00 (n/a)24594.30 (n/a)24542.14 (n/a)24551.90 (n/a)24495.50 (n/a)37.80 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:001.03 (+0.05%)1.03 (+0.22%)1.03 (+0.01%)1.03 (+0.88%)0.00 (-87.39%)24533.80 (-0.87%)24524.60 (-0.22%)24529.30 (-0.01%)24502.70 (-0.05%)12.46 (-87.49%)701.14 (+0.05%)700.52 (+0.22%)700.38 (+0.01%)700.25 (+0.88%)0.36 (-87.39%)
ea7f586 — 2025-12-18 22:24:431.03 (n/a)1.02 (n/a)1.03 (n/a)1.02 (n/a)0.00 (n/a)24749.90 (n/a)24579.28 (n/a)24531.30 (n/a)24514.30 (n/a)99.64 (n/a)700.81 (n/a)698.97 (n/a)700.32 (n/a)694.14 (n/a)2.82 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.99 (+1.69%)0.97 (+0.29%)0.97 (-0.04%)0.97 (-0.15%)0.01 (+434.92%)26025.30 (+0.15%)25839.34 (-0.29%)25916.70 (+0.04%)25446.00 (-1.67%)202.05 (+425.44%)
0a6c11c — 2025-12-04 00:39:100.97 (n/a)0.97 (n/a)0.97 (n/a)0.97 (n/a)0.00 (n/a)25986.90 (n/a)25913.52 (n/a)25905.70 (n/a)25877.00 (n/a)38.45 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:000.97 (-0.04%)0.97 (-0.01%)0.97 (+0.02%)0.97 (+0.03%)0.00 (-41.24%)25916.10 (-0.03%)25903.50 (+0.01%)25901.10 (-0.02%)25886.70 (+0.04%)12.08 (-41.13%)663.66 (-0.04%)663.23 (-0.01%)663.29 (+0.02%)662.90 (+0.03%)0.31 (-41.24%)
ea7f586 — 2025-12-18 22:24:430.97 (n/a)0.97 (n/a)0.97 (n/a)0.97 (n/a)0.00 (n/a)25923.60 (n/a)25900.86 (n/a)25905.90 (n/a)25877.60 (n/a)20.53 (n/a)663.89 (n/a)663.29 (n/a)663.17 (n/a)662.71 (n/a)0.53 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.92 (-1.02%)0.92 (-0.66%)0.92 (-0.10%)0.91 (-0.78%)0.01 (-4.54%)27751.40 (+0.79%)27480.74 (+0.66%)27354.80 (+0.10%)27320.50 (+1.03%)182.69 (-2.84%)
0a6c11c — 2025-12-04 00:39:100.93 (n/a)0.92 (n/a)0.92 (n/a)0.91 (n/a)0.01 (n/a)27534.80 (n/a)27299.32 (n/a)27326.70 (n/a)27042.30 (n/a)188.04 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:000.92 (-0.18%)0.91 (-0.07%)0.91 (+0.23%)0.90 (-0.32%)0.00 (+31.25%)27811.60 (+0.32%)27627.92 (+0.07%)27558.40 (-0.23%)27474.40 (+0.18%)149.39 (+32.02%)625.30 (-0.18%)621.84 (-0.07%)623.40 (+0.23%)617.72 (-0.32%)3.36 (+31.25%)
ea7f586 — 2025-12-18 22:24:430.92 (n/a)0.91 (n/a)0.91 (n/a)0.91 (n/a)0.00 (n/a)27723.90 (n/a)27609.34 (n/a)27622.50 (n/a)27424.40 (n/a)113.16 (n/a)626.44 (n/a)622.26 (n/a)621.95 (n/a)619.68 (n/a)2.56 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)GFLOP/s (max)GFLOP/s (mean)GFLOP/s (median)GFLOP/s (min)GFLOP/s (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
31ab24e — 2025-11-22 02:01:563.65 (-1.06%)3.59 (+0.83%)3.63 (-0.26%)3.45 (+2.66%)0.07 (-44.71%)2678.08 (-1.06%)2632.13 (+0.83%)2658.60 (-0.26%)2528.31 (+2.66%)53.86 (-44.71%)6795.00 (-2.59%)6529.80 (-0.92%)6462.00 (+0.26%)6415.00 (+1.07%)137.14 (-45.27%)
d9a64c5 — 2025-11-21 20:49:353.69 (n/a)3.56 (n/a)3.64 (n/a)3.36 (n/a)0.13 (n/a)2706.77 (n/a)2610.50 (n/a)2665.61 (n/a)2462.71 (n/a)97.41 (n/a)6976.00 (n/a)6590.40 (n/a)6445.00 (n/a)6347.00 (n/a)250.58 (n/a)

layer_norm_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-8.34%)0.06 (+2.58%)0.06 (+5.63%)0.05 (+16.32%)0.00 (-48.86%)156.10 (-14.04%)145.30 (-3.90%)144.60 (-5.30%)128.00 (+9.12%)11.19 (-50.94%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)181.60 (n/a)151.20 (n/a)152.70 (n/a)117.30 (n/a)22.81 (n/a)

layer_norm_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.09 (+16.24%)0.06 (+0.07%)0.06 (+1.13%)0.04 (-7.09%)0.02 (+43.89%)184.10 (+7.60%)140.68 (+2.00%)138.80 (-1.14%)95.00 (-14.03%)31.93 (+30.91%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)171.10 (n/a)137.92 (n/a)140.40 (n/a)110.50 (n/a)24.39 (n/a)

layer_norm_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-7.32%)0.06 (+2.87%)0.06 (+5.33%)0.05 (+12.01%)0.01 (-35.11%)163.30 (-10.72%)142.30 (-4.24%)141.20 (-5.11%)120.10 (+7.91%)16.41 (-36.23%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)182.90 (n/a)148.60 (n/a)148.80 (n/a)111.30 (n/a)25.73 (n/a)

layer_norm_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (+15.75%)0.06 (+5.69%)0.06 (+3.09%)0.03 (-1.95%)0.01 (+22.99%)234.80 (+2.00%)157.54 (-3.90%)147.00 (-3.03%)113.80 (-13.66%)46.27 (+13.76%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)230.20 (n/a)163.94 (n/a)151.60 (n/a)131.80 (n/a)40.67 (n/a)

layer_norm_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (-7.94%)0.05 (-17.05%)0.05 (-15.48%)0.04 (-29.61%)0.01 (+120.00%)230.10 (+42.04%)184.64 (+23.41%)179.10 (+18.30%)149.70 (+8.64%)34.35 (+238.08%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.05 (n/a)0.05 (n/a)0.05 (n/a)0.00 (n/a)162.00 (n/a)149.62 (n/a)151.40 (n/a)137.80 (n/a)10.16 (n/a)

layer_norm_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-6.31%)0.05 (+7.33%)0.06 (+10.43%)0.04 (+10.61%)0.01 (-26.01%)202.60 (-9.59%)157.18 (-8.47%)145.50 (-9.46%)132.20 (+6.79%)27.56 (-27.21%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)224.10 (n/a)171.72 (n/a)160.70 (n/a)123.80 (n/a)37.86 (n/a)

layer_norm_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-11.80%)0.05 (-13.81%)0.05 (-20.81%)0.04 (-1.40%)0.01 (-23.04%)227.60 (+1.38%)175.32 (+14.19%)178.10 (+26.22%)136.10 (+13.32%)34.29 (-15.94%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)224.50 (n/a)153.54 (n/a)141.10 (n/a)120.10 (n/a)40.79 (n/a)

layer_norm_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (-26.37%)0.04 (-6.91%)0.04 (+2.09%)0.04 (+21.11%)0.00 (-79.24%)197.80 (-17.45%)184.40 (+1.36%)185.90 (-2.05%)168.20 (+35.75%)11.36 (-76.50%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.05 (n/a)0.04 (n/a)0.03 (n/a)0.01 (n/a)239.60 (n/a)181.92 (n/a)189.80 (n/a)123.90 (n/a)48.32 (n/a)

matrix_vector_mul_128x128_32_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:490.37 (+17.13%)0.30 (+24.87%)0.30 (+38.03%)0.23 (+12.47%)0.06 (+35.23%)
0a6c11c — 2025-12-04 00:39:100.32 (n/a)0.24 (n/a)0.22 (n/a)0.20 (n/a)0.04 (n/a)

matrix_vector_mul_128x128_32_1col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:000.29 (+29.75%)0.25 (+22.55%)0.24 (+14.81%)0.22 (+20.59%)0.03 (+76.05%)0.29 (+29.75%)0.25 (+22.55%)0.24 (+14.81%)0.22 (+20.59%)0.03 (+76.04%)
ea7f586 — 2025-12-18 22:24:430.23 (n/a)0.21 (n/a)0.21 (n/a)0.18 (n/a)0.02 (n/a)0.22 (n/a)0.20 (n/a)0.21 (n/a)0.18 (n/a)0.02 (n/a)

matrix_vector_mul_2048x8192_1_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4913.57 (+3.47%)13.15 (+1.09%)13.10 (+0.16%)12.52 (-1.58%)0.37 (+153.19%)
0a6c11c — 2025-12-04 00:39:1013.11 (n/a)13.01 (n/a)13.07 (n/a)12.72 (n/a)0.15 (n/a)

matrix_vector_mul_2048x8192_1_1col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:0012.84 (-3.59%)12.45 (-1.98%)12.59 (-5.16%)11.77 (+6.47%)0.44 (-55.17%)12.83 (-3.59%)12.44 (-1.98%)12.58 (-5.16%)11.76 (+6.47%)0.44 (-55.17%)
ea7f586 — 2025-12-18 22:24:4313.31 (n/a)12.70 (n/a)13.28 (n/a)11.05 (n/a)0.97 (n/a)13.31 (n/a)12.70 (n/a)13.27 (n/a)11.04 (n/a)0.97 (n/a)

matrix_vector_mul_2048x8192_1_2col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4925.06 (+2.47%)23.71 (+1.81%)24.13 (+0.50%)20.90 (+5.41%)1.48 (-15.60%)
0a6c11c — 2025-12-04 00:39:1024.46 (n/a)23.29 (n/a)24.01 (n/a)19.83 (n/a)1.75 (n/a)

matrix_vector_mul_2048x8192_1_2col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:0025.04 (-0.74%)24.46 (+6.40%)24.38 (+2.10%)24.13 (+37.65%)0.37 (-88.04%)25.02 (-0.74%)24.45 (+6.40%)24.36 (+2.10%)24.11 (+37.65%)0.37 (-88.04%)
ea7f586 — 2025-12-18 22:24:4325.22 (n/a)22.99 (n/a)23.88 (n/a)17.53 (n/a)3.11 (n/a)25.21 (n/a)22.98 (n/a)23.86 (n/a)17.52 (n/a)3.11 (n/a)

matrix_vector_mul_2048x8192_1_4col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4942.05 (+1.54%)40.50 (+12.60%)40.49 (+2.36%)38.98 (+97.53%)0.97 (-88.09%)
0a6c11c — 2025-12-04 00:39:1041.41 (n/a)35.97 (n/a)39.55 (n/a)19.73 (n/a)8.15 (n/a)

matrix_vector_mul_2048x8192_1_4col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:0041.20 (-1.12%)39.57 (+9.35%)39.30 (-1.39%)38.79 (+102.39%)0.94 (-90.16%)41.18 (-1.12%)39.55 (+9.35%)39.28 (-1.39%)38.77 (+102.39%)0.94 (-90.16%)
ea7f586 — 2025-12-18 22:24:4341.67 (n/a)36.19 (n/a)39.86 (n/a)19.17 (n/a)9.55 (n/a)41.64 (n/a)36.17 (n/a)39.83 (n/a)19.16 (n/a)9.54 (n/a)

matrix_vector_mul_2048x8192_1_8col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4946.62 (+3.11%)44.82 (+0.17%)44.78 (+0.17%)42.60 (-3.80%)1.60 (+367.72%)
0a6c11c — 2025-12-04 00:39:1045.21 (n/a)44.75 (n/a)44.70 (n/a)44.29 (n/a)0.34 (n/a)

matrix_vector_mul_2048x8192_1_8col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:0048.30 (+8.09%)41.60 (-5.10%)44.18 (+1.01%)26.81 (-37.68%)8.45 (+1092.72%)48.27 (+8.09%)41.57 (-5.10%)44.15 (+1.01%)26.80 (-37.68%)8.45 (+1092.72%)
ea7f586 — 2025-12-18 22:24:4344.68 (n/a)43.83 (n/a)43.74 (n/a)43.02 (n/a)0.71 (n/a)44.66 (n/a)43.80 (n/a)43.71 (n/a)43.00 (n/a)0.71 (n/a)

matrix_vector_mul_8192x2048_4_1col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4913.29 (+1.63%)13.19 (+6.08%)13.19 (+1.17%)13.09 (+26.86%)0.08 (-92.94%)
0a6c11c — 2025-12-04 00:39:1013.08 (n/a)12.44 (n/a)13.03 (n/a)10.32 (n/a)1.07 (n/a)

matrix_vector_mul_8192x2048_4_1col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:0013.34 (+0.61%)12.03 (-3.63%)11.34 (-14.00%)11.04 (-0.94%)1.19 (+16.08%)13.33 (+0.61%)12.02 (-3.63%)11.34 (-14.00%)11.03 (-0.94%)1.19 (+16.08%)
ea7f586 — 2025-12-18 22:24:4313.26 (n/a)12.48 (n/a)13.19 (n/a)11.14 (n/a)1.03 (n/a)13.25 (n/a)12.47 (n/a)13.18 (n/a)11.14 (n/a)1.03 (n/a)

matrix_vector_mul_8192x2048_4_2col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4925.58 (+4.24%)24.23 (+6.59%)24.03 (+0.29%)23.58 (+37.29%)0.70 (-74.97%)
0a6c11c — 2025-12-04 00:39:1024.54 (n/a)22.73 (n/a)23.96 (n/a)17.18 (n/a)2.79 (n/a)

matrix_vector_mul_8192x2048_4_2col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:0024.88 (-0.81%)24.33 (-0.03%)24.33 (+0.42%)23.76 (+0.29%)0.40 (-22.92%)24.87 (-0.81%)24.31 (-0.03%)24.31 (+0.42%)23.74 (+0.29%)0.40 (-22.92%)
ea7f586 — 2025-12-18 22:24:4325.09 (n/a)24.34 (n/a)24.23 (n/a)23.69 (n/a)0.52 (n/a)25.07 (n/a)24.32 (n/a)24.21 (n/a)23.67 (n/a)0.52 (n/a)

matrix_vector_mul_8192x2048_4_4col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4940.94 (-2.19%)39.74 (-1.10%)39.76 (-0.60%)38.17 (-2.14%)1.04 (+4.39%)
0a6c11c — 2025-12-04 00:39:1041.86 (n/a)40.18 (n/a)40.00 (n/a)39.00 (n/a)1.00 (n/a)

matrix_vector_mul_8192x2048_4_4col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:0041.37 (+1.72%)39.82 (+3.09%)39.78 (-0.61%)38.27 (+10.49%)1.30 (-47.70%)41.35 (+1.72%)39.80 (+3.09%)39.75 (-0.61%)38.24 (+10.49%)1.30 (-47.70%)
ea7f586 — 2025-12-18 22:24:4340.67 (n/a)38.63 (n/a)40.02 (n/a)34.63 (n/a)2.49 (n/a)40.65 (n/a)38.60 (n/a)40.00 (n/a)34.61 (n/a)2.49 (n/a)

matrix_vector_mul_8192x2048_4_8col

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)
130b6ea — 2025-12-05 21:15:4946.71 (-1.04%)44.91 (-2.34%)44.71 (-2.98%)43.80 (-1.44%)0.96 (+6.93%)
0a6c11c — 2025-12-04 00:39:1047.20 (n/a)45.98 (n/a)46.08 (n/a)44.44 (n/a)0.90 (n/a)

matrix_vector_mul_8192x2048_4_8col0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Throughput (max)Throughput (mean)Throughput (median)Throughput (min)Throughput (stddev)
b7882b4 — 2025-12-18 23:16:0045.11 (-2.49%)44.03 (-0.71%)43.71 (-0.79%)43.52 (+0.67%)0.66 (-44.37%)45.08 (-2.49%)44.01 (-0.71%)43.68 (-0.79%)43.49 (+0.67%)0.66 (-44.37%)
ea7f586 — 2025-12-18 22:24:4346.26 (n/a)44.35 (n/a)44.05 (n/a)43.23 (n/a)1.19 (n/a)46.23 (n/a)44.32 (n/a)44.03 (n/a)43.20 (n/a)1.19 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.07 (+22.03%)0.06 (+27.86%)0.06 (+35.49%)0.05 (+15.00%)0.01 (+49.69%)173.30 (-13.00%)137.94 (-21.32%)133.60 (-26.19%)123.40 (-18.06%)18.40 (+8.53%)
0a6c11c — 2025-12-04 00:39:100.05 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.00 (n/a)199.20 (n/a)175.32 (n/a)181.00 (n/a)150.60 (n/a)16.96 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (+23.35%)0.05 (+35.03%)0.05 (+45.41%)0.04 (+28.70%)0.01 (-5.68%)189.50 (-22.30%)165.76 (-26.43%)164.60 (-31.24%)143.30 (-18.90%)16.97 (-40.38%)
ea7f586 — 2025-12-18 22:24:430.05 (n/a)0.04 (n/a)0.03 (n/a)0.03 (n/a)0.01 (n/a)243.90 (n/a)225.30 (n/a)239.40 (n/a)176.70 (n/a)28.46 (n/a)

mem_copy_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (-16.99%)0.05 (-12.47%)0.05 (-7.32%)0.04 (-11.17%)0.00 (-29.61%)197.00 (+12.57%)171.00 (+13.70%)164.00 (+7.89%)153.00 (+20.47%)17.49 (-4.32%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)175.00 (n/a)150.40 (n/a)152.00 (n/a)127.00 (n/a)18.28 (n/a)

mem_copy_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (+5.56%)0.05 (-1.59%)0.05 (-1.39%)0.03 (-14.07%)0.01 (+25.03%)263.00 (+16.37%)166.00 (+4.53%)144.00 (+1.41%)126.00 (-5.26%)49.38 (+43.17%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)226.00 (n/a)158.80 (n/a)142.00 (n/a)133.00 (n/a)34.49 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.08 (+10.93%)0.07 (+7.79%)0.07 (+3.55%)0.05 (-1.73%)0.01 (+36.73%)150.20 (+1.76%)121.68 (-6.36%)125.60 (-3.38%)97.90 (-9.85%)18.56 (+21.98%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.06 (n/a)0.06 (n/a)0.06 (n/a)0.01 (n/a)147.60 (n/a)129.94 (n/a)130.00 (n/a)108.60 (n/a)15.21 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.08 (+12.67%)0.06 (+4.49%)0.06 (+7.62%)0.05 (+1.80%)0.01 (+17.29%)162.20 (-1.76%)136.88 (-4.01%)140.00 (-7.04%)107.90 (-11.19%)19.55 (+2.62%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)165.10 (n/a)142.60 (n/a)150.60 (n/a)121.50 (n/a)19.05 (n/a)

mem_copy_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (-15.32%)0.05 (-17.09%)0.05 (-23.81%)0.04 (-15.54%)0.01 (-18.18%)193.00 (+18.40%)160.80 (+20.36%)168.00 (+31.25%)124.00 (+18.10%)22.92 (+11.39%)
0999a9a — 2025-11-13 19:24:080.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)163.00 (n/a)133.60 (n/a)128.00 (n/a)105.00 (n/a)20.58 (n/a)

mem_copy_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (-23.02%)0.05 (-6.98%)0.05 (-2.13%)0.04 (-15.31%)0.01 (-33.29%)196.00 (+18.07%)147.00 (+6.52%)141.00 (+2.17%)126.00 (+29.90%)25.27 (+8.13%)
0999a9a — 2025-11-13 19:24:080.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)166.00 (n/a)138.00 (n/a)138.00 (n/a)97.00 (n/a)23.37 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.10 (+31.37%)0.08 (+29.12%)0.08 (+29.06%)0.07 (+41.80%)0.01 (+0.23%)109.90 (-29.46%)98.30 (-23.38%)100.80 (-22.52%)79.40 (-23.87%)10.54 (-46.17%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.07 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)155.80 (n/a)128.30 (n/a)130.10 (n/a)104.30 (n/a)19.58 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.08 (+11.52%)0.06 (-1.33%)0.05 (-4.36%)0.03 (-35.27%)0.02 (+87.13%)281.90 (+54.47%)163.34 (+11.42%)153.30 (+4.50%)99.20 (-10.31%)70.36 (+176.28%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)182.50 (n/a)146.60 (n/a)146.70 (n/a)110.60 (n/a)25.47 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.10 (+25.02%)0.07 (+10.01%)0.07 (+7.53%)0.05 (-1.99%)0.01 (+77.08%)150.10 (+2.04%)114.60 (-7.22%)111.70 (-6.99%)83.90 (-20.02%)21.23 (+43.54%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)147.10 (n/a)123.52 (n/a)120.10 (n/a)104.90 (n/a)14.79 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-0.26%)0.06 (+0.24%)0.07 (+18.49%)0.04 (-22.62%)0.01 (+74.90%)203.10 (+29.20%)142.36 (+3.26%)120.70 (-15.65%)114.20 (+0.26%)37.33 (+129.18%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)157.20 (n/a)137.86 (n/a)143.10 (n/a)113.90 (n/a)16.29 (n/a)

mem_copy_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.07 (+11.97%)0.05 (+3.09%)0.05 (+2.10%)0.03 (-11.35%)0.01 (+41.65%)229.00 (+12.81%)156.00 (-0.51%)143.00 (-2.05%)117.00 (-10.69%)38.21 (+48.71%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)203.00 (n/a)156.80 (n/a)146.00 (n/a)131.00 (n/a)25.69 (n/a)

mem_copy_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.06 (+9.02%)0.05 (+5.59%)0.05 (+0.00%)0.04 (+3.63%)0.01 (+25.48%)193.00 (-3.50%)147.40 (-4.66%)147.00 (+0.00%)122.00 (-8.27%)25.24 (+7.48%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)200.00 (n/a)154.60 (n/a)147.00 (n/a)133.00 (n/a)23.48 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.09 (-3.98%)0.07 (+4.00%)0.07 (+7.04%)0.06 (+3.67%)0.01 (-15.76%)129.70 (-3.57%)113.30 (-4.50%)116.90 (-6.55%)90.40 (+4.15%)15.04 (-13.80%)
0a6c11c — 2025-12-04 00:39:100.09 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)134.50 (n/a)118.64 (n/a)125.10 (n/a)86.80 (n/a)17.45 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-16.98%)0.05 (-0.45%)0.05 (+5.35%)0.05 (+42.64%)0.00 (-86.42%)157.90 (-29.92%)152.50 (-4.18%)152.10 (-5.12%)146.80 (+20.53%)4.75 (-88.50%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)225.30 (n/a)159.16 (n/a)160.30 (n/a)121.80 (n/a)41.28 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.08 (+9.57%)0.07 (+16.85%)0.07 (+19.58%)0.06 (+24.52%)0.01 (+1.05%)134.50 (-19.70%)116.64 (-14.92%)116.80 (-16.39%)98.00 (-8.75%)15.71 (-25.13%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)167.50 (n/a)137.10 (n/a)139.70 (n/a)107.40 (n/a)20.99 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.08 (+6.86%)0.06 (-1.18%)0.06 (-4.35%)0.03 (-0.83%)0.02 (-4.18%)234.30 (+0.86%)151.50 (+0.34%)140.60 (+4.54%)99.20 (-6.42%)50.36 (-4.28%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.02 (n/a)232.30 (n/a)150.98 (n/a)134.50 (n/a)106.00 (n/a)52.61 (n/a)

mem_copy_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (-13.48%)0.05 (-17.79%)0.05 (-15.33%)0.03 (-25.09%)0.01 (+6.67%)267.00 (+33.50%)176.80 (+24.51%)150.00 (+18.11%)141.00 (+15.57%)47.41 (+61.41%)
0999a9a — 2025-11-13 19:24:080.06 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)200.00 (n/a)142.00 (n/a)127.00 (n/a)122.00 (n/a)29.37 (n/a)

mem_copy_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
003b49d — 2025-11-14 19:30:540.05 (+13.38%)0.05 (+23.85%)0.05 (+37.91%)0.05 (+30.72%)0.00 (-56.85%)166.00 (-23.50%)153.40 (-20.60%)153.00 (-27.49%)142.00 (-11.80%)7.63 (-70.60%)
0999a9a — 2025-11-13 19:24:080.05 (n/a)0.04 (n/a)0.04 (n/a)0.04 (n/a)0.01 (n/a)217.00 (n/a)193.20 (n/a)211.00 (n/a)161.00 (n/a)25.96 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.09 (+16.31%)0.08 (+15.34%)0.08 (+14.60%)0.07 (+19.30%)0.01 (-5.41%)117.30 (-16.15%)107.80 (-13.61%)108.50 (-12.78%)94.10 (-13.99%)8.29 (-32.01%)
0a6c11c — 2025-12-04 00:39:100.07 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)139.90 (n/a)124.78 (n/a)124.40 (n/a)109.40 (n/a)12.20 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (+1.09%)0.06 (+15.10%)0.06 (+21.78%)0.04 (+26.35%)0.01 (-24.66%)190.40 (-20.83%)149.40 (-15.04%)140.60 (-17.87%)126.40 (-1.10%)24.68 (-40.26%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.05 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)240.50 (n/a)175.84 (n/a)171.20 (n/a)127.80 (n/a)41.31 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.09 (+12.61%)0.07 (-2.24%)0.07 (-8.85%)0.06 (+5.13%)0.01 (+35.50%)136.50 (-4.88%)120.86 (+2.96%)125.10 (+9.64%)91.60 (-11.15%)15.72 (+10.46%)
0a6c11c — 2025-12-04 00:39:100.08 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)143.50 (n/a)117.38 (n/a)114.10 (n/a)103.10 (n/a)14.23 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (+8.01%)0.05 (+4.71%)0.05 (+14.73%)0.04 (-6.63%)0.01 (+64.20%)223.00 (+7.11%)178.42 (-2.20%)158.90 (-12.84%)138.30 (-7.43%)39.12 (+69.57%)
ea7f586 — 2025-12-18 22:24:430.05 (n/a)0.05 (n/a)0.04 (n/a)0.04 (n/a)0.01 (n/a)208.20 (n/a)182.44 (n/a)182.30 (n/a)149.40 (n/a)23.07 (n/a)

mha

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.18 (+0.16%)0.18 (+0.19%)0.18 (+0.12%)0.18 (+0.32%)0.00 (-52.24%)47253.90 (-0.32%)47213.70 (-0.19%)47210.90 (-0.12%)47167.70 (-0.16%)28.71 (-52.47%)
0a6c11c — 2025-12-04 00:39:100.18 (n/a)0.18 (n/a)0.18 (n/a)0.18 (n/a)0.00 (n/a)47405.00 (n/a)47301.72 (n/a)47265.90 (n/a)47241.80 (n/a)60.40 (n/a)

mha0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.18 (+0.66%)0.18 (+0.19%)0.18 (+0.09%)0.18 (-0.06%)0.00 (+99.87%)47500.00 (+0.06%)47196.86 (-0.19%)47197.10 (-0.09%)46916.40 (-0.65%)209.71 (+98.68%)
ea7f586 — 2025-12-18 22:24:430.18 (n/a)0.18 (n/a)0.18 (n/a)0.18 (n/a)0.00 (n/a)47473.20 (n/a)47284.94 (n/a)47240.60 (n/a)47224.50 (n/a)105.55 (n/a)

relu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-16.32%)0.05 (-13.02%)0.06 (-13.76%)0.05 (-11.16%)0.01 (-21.19%)179.00 (+12.58%)155.58 (+14.62%)146.90 (+15.94%)136.90 (+19.56%)21.07 (+4.30%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)159.00 (n/a)135.74 (n/a)126.70 (n/a)114.50 (n/a)20.21 (n/a)

relu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-1.90%)0.06 (-1.26%)0.06 (-1.58%)0.05 (+2.73%)0.00 (-18.40%)157.00 (-2.67%)139.72 (+1.00%)134.80 (+1.66%)128.80 (+1.98%)11.79 (-19.06%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)161.30 (n/a)138.34 (n/a)132.60 (n/a)126.30 (n/a)14.56 (n/a)

relu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-9.24%)0.06 (-6.90%)0.06 (-10.02%)0.05 (+1.21%)0.01 (-38.30%)169.60 (-1.17%)144.32 (+6.27%)137.50 (+11.16%)132.10 (+10.18%)15.02 (-31.58%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)171.60 (n/a)135.80 (n/a)123.70 (n/a)119.90 (n/a)21.95 (n/a)

relu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (+8.42%)0.05 (+8.50%)0.05 (+9.02%)0.05 (+11.06%)0.01 (+0.61%)179.80 (-9.96%)159.58 (-8.00%)163.10 (-8.27%)135.30 (-7.71%)18.42 (-15.99%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)199.70 (n/a)173.46 (n/a)177.80 (n/a)146.60 (n/a)21.92 (n/a)

rms_norm_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-16.20%)0.06 (-0.76%)0.06 (+6.36%)0.05 (+11.60%)0.01 (-42.34%)161.20 (-10.39%)137.74 (-1.32%)130.40 (-5.98%)121.20 (+19.29%)17.45 (-37.40%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)179.90 (n/a)139.58 (n/a)138.70 (n/a)101.60 (n/a)27.88 (n/a)

rms_norm_1_cols_2_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-7.22%)0.06 (-7.73%)0.05 (-11.48%)0.04 (-14.47%)0.01 (+10.64%)183.40 (+16.89%)150.34 (+8.93%)151.00 (+12.94%)127.10 (+7.71%)21.10 (+40.53%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)156.90 (n/a)138.02 (n/a)133.70 (n/a)118.00 (n/a)15.02 (n/a)

rms_norm_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-6.43%)0.06 (-5.53%)0.05 (-6.96%)0.05 (-4.36%)0.00 (-16.71%)163.30 (+4.55%)149.54 (+5.70%)152.20 (+7.49%)134.20 (+6.93%)12.04 (-7.82%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)156.20 (n/a)141.48 (n/a)141.60 (n/a)125.50 (n/a)13.07 (n/a)

rms_norm_2_cols_2_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-21.02%)0.06 (-17.34%)0.05 (-25.33%)0.04 (+14.32%)0.01 (-41.62%)184.50 (-12.52%)151.16 (+15.11%)159.00 (+33.95%)113.00 (+26.68%)27.59 (-40.41%)
ea7f586 — 2025-12-18 22:24:430.09 (n/a)0.07 (n/a)0.07 (n/a)0.04 (n/a)0.02 (n/a)210.90 (n/a)131.32 (n/a)118.70 (n/a)89.20 (n/a)46.31 (n/a)

rms_norm_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-35.32%)0.05 (-16.22%)0.05 (-10.56%)0.05 (-6.00%)0.00 (-78.68%)167.20 (+6.36%)157.42 (+15.55%)159.30 (+11.79%)143.70 (+54.52%)9.15 (-63.53%)
ea7f586 — 2025-12-18 22:24:430.09 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)157.20 (n/a)136.24 (n/a)142.50 (n/a)93.00 (n/a)25.08 (n/a)

rms_norm_4_cols_2_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-0.49%)0.06 (-6.80%)0.07 (-5.82%)0.05 (-18.28%)0.01 (+59.08%)175.30 (+22.42%)137.42 (+8.82%)125.90 (+6.16%)117.40 (+0.51%)23.49 (+99.08%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.07 (n/a)0.07 (n/a)0.06 (n/a)0.01 (n/a)143.20 (n/a)126.28 (n/a)118.60 (n/a)116.80 (n/a)11.80 (n/a)

rms_norm_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (+10.76%)0.05 (-12.00%)0.05 (-19.73%)0.04 (-0.33%)0.01 (+29.20%)233.60 (+0.30%)184.32 (+15.52%)181.50 (+24.57%)117.50 (-9.68%)45.74 (+10.14%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.05 (n/a)0.06 (n/a)0.04 (n/a)0.01 (n/a)232.90 (n/a)159.56 (n/a)145.70 (n/a)130.10 (n/a)41.53 (n/a)

rms_norm_8_cols_2_channels_2048_tile_128

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-13.31%)0.05 (-12.03%)0.05 (-20.05%)0.04 (-13.76%)0.01 (+4.46%)196.50 (+16.00%)163.00 (+14.63%)176.80 (+25.04%)129.80 (+15.38%)29.49 (+36.46%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)169.40 (n/a)142.20 (n/a)141.40 (n/a)112.50 (n/a)21.61 (n/a)

rope_1_cols_2_channels_4096_tile_4096_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.25 (+8.71%)0.20 (+7.01%)0.21 (+10.18%)0.16 (+0.56%)0.03 (+28.53%)150.90 (-0.53%)122.54 (-5.91%)119.80 (-9.31%)98.90 (-8.00%)19.36 (+19.23%)
ea7f586 — 2025-12-18 22:24:430.23 (n/a)0.19 (n/a)0.19 (n/a)0.16 (n/a)0.02 (n/a)151.70 (n/a)130.24 (n/a)132.10 (n/a)107.50 (n/a)16.24 (n/a)

rope_2_cols_2_channels_4096_tile_2048_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.22 (-3.02%)0.19 (-5.29%)0.19 (-6.65%)0.15 (-9.53%)0.03 (+21.40%)159.60 (+10.53%)133.10 (+6.31%)130.40 (+7.15%)112.70 (+3.11%)19.88 (+36.73%)
ea7f586 — 2025-12-18 22:24:430.22 (n/a)0.20 (n/a)0.20 (n/a)0.17 (n/a)0.02 (n/a)144.40 (n/a)125.20 (n/a)121.70 (n/a)109.30 (n/a)14.54 (n/a)

rope_4_cols_2_channels_4096_tile_1024_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.24 (+21.39%)0.21 (+21.16%)0.20 (+18.82%)0.18 (+27.28%)0.02 (+12.56%)133.00 (-21.44%)118.94 (-17.62%)122.50 (-15.81%)101.90 (-17.62%)12.31 (-27.62%)
ea7f586 — 2025-12-18 22:24:430.20 (n/a)0.17 (n/a)0.17 (n/a)0.15 (n/a)0.02 (n/a)169.30 (n/a)144.38 (n/a)145.50 (n/a)123.70 (n/a)17.01 (n/a)

rope_8_cols_2_channels_4096_tile_512_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.25 (+33.40%)0.18 (+9.17%)0.16 (+5.67%)0.15 (+0.95%)0.04 (+157.04%)161.90 (-0.92%)142.38 (-6.01%)150.70 (-5.34%)98.30 (-25.08%)25.33 (+84.02%)
ea7f586 — 2025-12-18 22:24:430.19 (n/a)0.16 (n/a)0.15 (n/a)0.15 (n/a)0.02 (n/a)163.40 (n/a)151.48 (n/a)159.20 (n/a)131.20 (n/a)13.76 (n/a)

sigmoid_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-5.36%)0.06 (+1.21%)0.06 (+3.51%)0.05 (+6.76%)0.01 (-29.90%)150.90 (-6.33%)138.70 (-1.94%)139.70 (-3.39%)119.60 (+5.65%)13.05 (-28.98%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)161.10 (n/a)141.44 (n/a)144.60 (n/a)113.20 (n/a)18.38 (n/a)

sigmoid_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-8.77%)0.06 (-6.23%)0.06 (-12.01%)0.05 (-1.71%)0.01 (-30.13%)155.50 (+1.70%)133.90 (+5.33%)132.00 (+13.60%)112.40 (+9.55%)17.44 (-24.49%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)152.90 (n/a)127.12 (n/a)116.20 (n/a)102.60 (n/a)23.10 (n/a)

sigmoid_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (+7.89%)0.06 (-0.20%)0.06 (-2.40%)0.05 (-2.22%)0.01 (+52.99%)164.00 (+2.31%)144.22 (+1.07%)148.30 (+2.42%)115.30 (-7.32%)19.58 (+45.48%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)160.30 (n/a)142.70 (n/a)144.80 (n/a)124.40 (n/a)13.46 (n/a)

sigmoid_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.08 (-10.45%)0.06 (-5.69%)0.06 (+1.28%)0.05 (-6.09%)0.01 (-14.62%)178.20 (+6.45%)139.48 (+5.35%)139.00 (-1.28%)104.70 (+11.62%)31.79 (+1.53%)
ea7f586 — 2025-12-18 22:24:430.09 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.02 (n/a)167.40 (n/a)132.40 (n/a)140.80 (n/a)93.80 (n/a)31.31 (n/a)

silu_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-5.20%)0.06 (+14.35%)0.06 (+20.24%)0.05 (+23.23%)0.01 (-35.19%)166.00 (-18.83%)140.46 (-14.76%)135.80 (-16.84%)117.80 (+5.46%)20.01 (-42.08%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.05 (n/a)0.05 (n/a)0.04 (n/a)0.01 (n/a)204.50 (n/a)164.78 (n/a)163.30 (n/a)111.70 (n/a)34.54 (n/a)

silu_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (+4.48%)0.05 (-8.83%)0.05 (-13.01%)0.04 (-21.46%)0.01 (+156.91%)193.40 (+27.32%)155.06 (+12.95%)150.20 (+15.01%)123.90 (-4.32%)31.81 (+211.54%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.00 (n/a)151.90 (n/a)137.28 (n/a)130.60 (n/a)129.50 (n/a)10.21 (n/a)

silu_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-1.17%)0.05 (-5.27%)0.06 (+14.81%)0.03 (-27.60%)0.01 (+19.26%)236.80 (+38.16%)162.18 (+8.50%)146.00 (-12.94%)114.00 (+1.15%)46.04 (+68.49%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.05 (n/a)0.05 (n/a)0.01 (n/a)171.40 (n/a)149.48 (n/a)167.70 (n/a)112.70 (n/a)27.33 (n/a)

silu_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-26.13%)0.05 (-10.38%)0.05 (-4.17%)0.04 (+16.58%)0.01 (-53.29%)206.60 (-14.20%)163.56 (+5.50%)155.60 (+4.36%)141.10 (+35.28%)26.90 (-47.90%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.05 (n/a)0.03 (n/a)0.02 (n/a)240.80 (n/a)155.04 (n/a)149.10 (n/a)104.30 (n/a)51.64 (n/a)

softmax_1_cols_2_channels_4096_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.15 (-4.44%)0.13 (+7.15%)0.14 (+25.48%)0.08 (+1.50%)0.03 (+0.31%)202.30 (-1.46%)135.68 (-6.40%)115.40 (-20.30%)106.90 (+4.60%)39.46 (+3.32%)
ea7f586 — 2025-12-18 22:24:430.16 (n/a)0.12 (n/a)0.11 (n/a)0.08 (n/a)0.03 (n/a)205.30 (n/a)144.96 (n/a)144.80 (n/a)102.20 (n/a)38.19 (n/a)

softmax_2_cols_2_channels_4096_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.15 (+7.00%)0.13 (+17.51%)0.14 (+27.32%)0.10 (+27.43%)0.02 (-5.02%)172.10 (-21.52%)128.52 (-16.21%)114.70 (-21.44%)107.10 (-6.54%)26.61 (-32.49%)
ea7f586 — 2025-12-18 22:24:430.14 (n/a)0.11 (n/a)0.11 (n/a)0.07 (n/a)0.02 (n/a)219.30 (n/a)153.38 (n/a)146.00 (n/a)114.60 (n/a)39.42 (n/a)

softmax_2_cols_2_channels_4096_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.17 (+10.51%)0.14 (+6.35%)0.14 (+5.26%)0.07 (-21.72%)0.04 (+37.46%)229.20 (+27.76%)133.34 (-0.83%)115.90 (-4.92%)94.50 (-9.57%)54.46 (+70.96%)
ea7f586 — 2025-12-18 22:24:430.16 (n/a)0.13 (n/a)0.13 (n/a)0.09 (n/a)0.03 (n/a)179.40 (n/a)134.46 (n/a)121.90 (n/a)104.50 (n/a)31.86 (n/a)

swiglu

No metrics available.

swiglu_decode_1x2048x2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:490.00 (-15.46%)0.00 (-27.20%)0.00 (-28.17%)0.00 (-33.90%)0.00 (+1834.83%)5187.20 (+51.28%)4720.84 (+38.38%)4750.90 (+39.22%)4015.10 (+18.29%)390.23 (+3298.45%)
0a6c11c — 2025-12-04 00:39:100.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)3428.80 (n/a)3411.62 (n/a)3412.60 (n/a)3394.20 (n/a)11.48 (n/a)

swiglu_decode_1x2048x2048_0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.00 (-0.66%)0.00 (+3.51%)0.00 (+7.74%)0.00 (+10.57%)0.00 (-14.22%)5042.80 (-9.56%)4533.84 (-4.28%)4675.10 (-7.19%)3414.30 (+0.67%)659.58 (-20.58%)
ea7f586 — 2025-12-18 22:24:430.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)0.00 (n/a)5575.90 (n/a)4736.54 (n/a)5037.10 (n/a)3391.70 (n/a)830.45 (n/a)

tanh_1_cols_1_channels_2048_tile_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.08 (+23.60%)0.06 (+3.36%)0.06 (-4.02%)0.06 (-1.25%)0.01 (+228.60%)147.90 (+1.23%)133.72 (-1.98%)140.00 (+4.17%)104.70 (-19.09%)16.87 (+160.52%)
ea7f586 — 2025-12-18 22:24:430.06 (n/a)0.06 (n/a)0.06 (n/a)0.06 (n/a)0.00 (n/a)146.10 (n/a)136.42 (n/a)134.40 (n/a)129.40 (n/a)6.48 (n/a)

tanh_2_cols_1_channels_2048_tile_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-30.78%)0.05 (-22.86%)0.05 (-25.24%)0.05 (-5.85%)0.00 (-75.60%)180.30 (+6.18%)161.68 (+24.77%)159.70 (+33.75%)148.10 (+44.49%)11.70 (-61.53%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.07 (n/a)0.07 (n/a)0.05 (n/a)0.01 (n/a)169.80 (n/a)129.58 (n/a)119.40 (n/a)102.50 (n/a)30.41 (n/a)

tanh_4_cols_1_channels_2048_tile_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (-33.47%)0.06 (-19.50%)0.06 (-19.71%)0.05 (+32.73%)0.00 (-75.82%)168.40 (-24.65%)148.56 (+13.77%)143.00 (+24.56%)137.60 (+50.38%)13.28 (-74.88%)
ea7f586 — 2025-12-18 22:24:430.09 (n/a)0.07 (n/a)0.07 (n/a)0.04 (n/a)0.02 (n/a)223.50 (n/a)130.58 (n/a)114.80 (n/a)91.50 (n/a)52.85 (n/a)

tanh_8_cols_1_channels_2048_tile_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.05 (-29.28%)0.04 (-30.73%)0.04 (-26.46%)0.03 (-41.66%)0.01 (+21.54%)272.70 (+71.40%)208.80 (+47.42%)193.10 (+35.99%)170.60 (+41.34%)41.69 (+199.94%)
ea7f586 — 2025-12-18 22:24:430.07 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)159.10 (n/a)141.64 (n/a)142.00 (n/a)120.70 (n/a)13.90 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:493.93 (-7.02%)3.68 (+2.48%)3.66 (+1.82%)3.45 (+15.62%)0.15 (-63.64%)151.80 (-13.50%)142.70 (-3.59%)143.40 (-1.78%)133.50 (+7.49%)5.90 (-66.22%)
0a6c11c — 2025-12-04 00:39:104.22 (n/a)3.59 (n/a)3.59 (n/a)2.99 (n/a)0.42 (n/a)175.50 (n/a)148.02 (n/a)146.00 (n/a)124.20 (n/a)17.46 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:004.00 (-3.23%)3.81 (+7.07%)3.86 (+5.40%)3.58 (+25.94%)0.16 (-71.77%)146.50 (-20.55%)137.76 (-8.53%)135.70 (-5.17%)131.20 (+3.39%)6.03 (-76.44%)
ea7f586 — 2025-12-18 22:24:434.13 (n/a)3.56 (n/a)3.66 (n/a)2.84 (n/a)0.58 (n/a)184.40 (n/a)150.60 (n/a)143.10 (n/a)126.90 (n/a)25.60 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
130b6ea — 2025-12-05 21:15:495.11 (+15.57%)3.85 (-3.32%)3.40 (-18.64%)3.29 (+7.09%)0.71 (+49.14%)159.60 (-6.61%)140.18 (+4.80%)154.20 (+22.97%)102.60 (-13.49%)22.70 (+19.77%)
0a6c11c — 2025-12-04 00:39:104.42 (n/a)3.99 (n/a)4.18 (n/a)3.07 (n/a)0.48 (n/a)170.90 (n/a)133.76 (n/a)125.40 (n/a)118.60 (n/a)18.95 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:004.93 (+11.43%)4.43 (+9.45%)4.49 (+5.69%)3.96 (+13.63%)0.43 (+2.29%)132.50 (-12.02%)119.24 (-8.77%)116.80 (-5.35%)106.30 (-10.22%)11.53 (-18.39%)
ea7f586 — 2025-12-18 22:24:434.43 (n/a)4.05 (n/a)4.25 (n/a)3.48 (n/a)0.42 (n/a)150.60 (n/a)130.70 (n/a)123.40 (n/a)118.40 (n/a)14.13 (n/a)

weighted_rms_norm_1_cols_2_channels_2048_weights_2048

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.07 (-13.65%)0.06 (-13.42%)0.05 (-13.91%)0.05 (-0.56%)0.01 (-37.02%)180.90 (+0.56%)150.38 (+13.24%)149.90 (+16.20%)121.50 (+15.82%)22.22 (-26.30%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)179.90 (n/a)132.80 (n/a)129.00 (n/a)104.90 (n/a)30.15 (n/a)

weighted_rms_norm_2_cols_2_channels_2048_weights_1024

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.08 (-1.51%)0.07 (+12.62%)0.08 (+37.18%)0.05 (+17.42%)0.01 (-16.92%)161.40 (-14.83%)122.40 (-12.97%)107.90 (-27.09%)98.30 (+1.55%)26.19 (-26.64%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.06 (n/a)0.04 (n/a)0.02 (n/a)189.50 (n/a)140.64 (n/a)148.00 (n/a)96.80 (n/a)35.70 (n/a)

weighted_rms_norm_4_cols_2_channels_2048_weights_512

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.08 (-3.58%)0.07 (+5.81%)0.07 (+18.45%)0.05 (+6.70%)0.01 (-29.76%)149.10 (-6.29%)121.90 (-6.93%)118.50 (-15.54%)102.10 (+3.65%)17.22 (-29.87%)
ea7f586 — 2025-12-18 22:24:430.08 (n/a)0.06 (n/a)0.06 (n/a)0.05 (n/a)0.01 (n/a)159.10 (n/a)130.98 (n/a)140.30 (n/a)98.50 (n/a)24.56 (n/a)

weighted_rms_norm_8_cols_2_channels_2048_weights_256

Commit/Date Bandwidth (max)Bandwidth (mean)Bandwidth (median)Bandwidth (min)Bandwidth (stddev)Latency (max)Latency (mean)Latency (median)Latency (min)Latency (stddev)
b7882b4 — 2025-12-18 23:16:000.06 (+10.88%)0.05 (+19.51%)0.05 (+16.07%)0.05 (+44.40%)0.00 (-57.84%)165.60 (-30.74%)153.90 (-17.99%)149.60 (-13.82%)144.60 (-9.79%)8.87 (-73.33%)
ea7f586 — 2025-12-18 22:24:430.05 (n/a)0.04 (n/a)0.05 (n/a)0.03 (n/a)0.01 (n/a)239.10 (n/a)187.66 (n/a)173.60 (n/a)160.30 (n/a)33.25 (n/a)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants