Add UInt16 argument intrinsics

[christiangnrd]

Implemented while debugging a proper fix for #719 but this is also a standalone contribution.

Not added to docstrings yet because similar to how there are limitations to mixing different vector sizes in the indexing intrinsics, there are also limitations to mixing intrinsic return types. See Section 5.2.3.6 of the Metal Shading Language Specification for details.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.19%. Comparing base (67d668c) to head (fddfd2e).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #725      +/-   ##
==========================================
+ Coverage   80.90%   81.19%   +0.29%     
==========================================
  Files          59       62       +3     
  Lines        2896     2904       +8     
==========================================
+ Hits         2343     2358      +15     
+ Misses        553      546       -7     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

Metal Benchmarks

Details

Benchmark suite	Current: fddfd2e	Previous: 67d668c	Ratio
latency/precompile	25183586750 ns	24820843000 ns	1.01
latency/ttfp	2300877084 ns	2257593833 ns	1.02
latency/import	1459733584 ns	1431203750 ns	1.02
integration/metaldevrt	842395.5 ns	834875 ns	1.01
integration/byval/slices=1	1580958 ns	1525666.5 ns	1.04
integration/byval/slices=3	19261562.5 ns	8498958 ns	2.27
integration/byval/reference	1578708 ns	1538166 ns	1.03
integration/byval/slices=2	2670750 ns	2552562 ns	1.05
kernel/indexing	472458 ns	593833 ns	0.80
kernel/indexing_checked	502916.5 ns	575750 ns	0.87
kernel/launch	12417 ns	11250 ns	1.10
kernel/rand	532416.5 ns	557187.5 ns	0.96
array/construct	6333 ns	6000 ns	1.06
array/broadcast	561270.5 ns	591209 ns	0.95
array/random/randn/Float32	898292 ns	836917 ns	1.07
array/random/randn!/Float32	594042 ns	619542 ns	0.96
array/random/rand!/Int64	551583 ns	548834 ns	1.01
array/random/rand!/Float32	542354.5 ns	593333 ns	0.91
array/random/rand/Int64	878375 ns	735667 ns	1.19
array/random/rand/Float32	840104.5 ns	631792 ns	1.33
array/accumulate/Int64/1d	1312625 ns	1237125 ns	1.06
array/accumulate/Int64/dims=1	1842812 ns	1795625 ns	1.03
array/accumulate/Int64/dims=2	2248209 ns	2130458 ns	1.06
array/accumulate/Int64/dims=1L	12092500 ns	11609562.5 ns	1.04
array/accumulate/Int64/dims=2L	9940292 ns	9610834 ns	1.03
array/accumulate/Float32/1d	1089125.5 ns	1111187.5 ns	0.98
array/accumulate/Float32/dims=1	1585375 ns	1518146 ns	1.04
array/accumulate/Float32/dims=2	1992375 ns	1836167 ns	1.09
array/accumulate/Float32/dims=1L	10231542 ns	9757375 ns	1.05
array/accumulate/Float32/dims=2L	7492709 ns	7203562.5 ns	1.04
array/reductions/reduce/Int64/1d	1333875 ns	1498333 ns	0.89
array/reductions/reduce/Int64/dims=1	1116938 ns	1076542 ns	1.04
array/reductions/reduce/Int64/dims=2	1151583 ns	1129417 ns	1.02
array/reductions/reduce/Int64/dims=1L	2042312.5 ns	2002083.5 ns	1.02
array/reductions/reduce/Int64/dims=2L	3920000 ns	4214895.5 ns	0.93
array/reductions/reduce/Float32/1d	767125 ns	991375 ns	0.77
array/reductions/reduce/Float32/dims=1	801271 ns	827000 ns	0.97
array/reductions/reduce/Float32/dims=2	822084 ns	833917 ns	0.99
array/reductions/reduce/Float32/dims=1L	1331792 ns	1305125 ns	1.02
array/reductions/reduce/Float32/dims=2L	1801375 ns	1788375 ns	1.01
array/reductions/mapreduce/Int64/1d	1319187.5 ns	1549292 ns	0.85
array/reductions/mapreduce/Int64/dims=1	1114541.5 ns	1085333 ns	1.03
array/reductions/mapreduce/Int64/dims=2	1160354 ns	1201959 ns	0.97
array/reductions/mapreduce/Int64/dims=1L	1990854 ns	2019583 ns	0.99
array/reductions/mapreduce/Int64/dims=2L	3620834 ns	3628521 ns	1.00
array/reductions/mapreduce/Float32/1d	847708 ns	1036542 ns	0.82
array/reductions/mapreduce/Float32/dims=1	796729 ns	819667 ns	0.97
array/reductions/mapreduce/Float32/dims=2	842667 ns	843917 ns	1.00
array/reductions/mapreduce/Float32/dims=1L	1348896 ns	1280500 ns	1.05
array/reductions/mapreduce/Float32/dims=2L	1804458 ns	1784500 ns	1.01
array/private/copyto!/gpu_to_gpu	556583 ns	635375 ns	0.88
array/private/copyto!/cpu_to_gpu	762375 ns	786625 ns	0.97
array/private/copyto!/gpu_to_cpu	700083.5 ns	773833 ns	0.90
array/private/iteration/findall/int	1585979.5 ns	1620458 ns	0.98
array/private/iteration/findall/bool	1478416 ns	1430125 ns	1.03
array/private/iteration/findfirst/int	2072000 ns	2024937.5 ns	1.02
array/private/iteration/findfirst/bool	2033792 ns	2010916 ns	1.01
array/private/iteration/scalar	3401500 ns	5600375 ns	0.61
array/private/iteration/logical	2683958 ns	2504521 ns	1.07
array/private/iteration/findmin/1d	2291938 ns	2209917 ns	1.04
array/private/iteration/findmin/2d	1542792 ns	1498584 ns	1.03
array/private/copy	833875 ns	558312.5 ns	1.49
array/shared/copyto!/gpu_to_gpu	85500 ns	82042 ns	1.04
array/shared/copyto!/cpu_to_gpu	85042 ns	79750 ns	1.07
array/shared/copyto!/gpu_to_cpu	84708 ns	82125 ns	1.03
array/shared/iteration/findall/int	1588895.5 ns	1600354 ns	0.99
array/shared/iteration/findall/bool	1493771 ns	1452458 ns	1.03
array/shared/iteration/findfirst/int	1694812.5 ns	1621520.5 ns	1.05
array/shared/iteration/findfirst/bool	1650687 ns	1607916.5 ns	1.03
array/shared/iteration/scalar	209708 ns	202916 ns	1.03
array/shared/iteration/logical	2301833 ns	2386416.5 ns	0.96
array/shared/iteration/findmin/1d	1902084 ns	1799396 ns	1.06
array/shared/iteration/findmin/2d	1539312.5 ns	1500416.5 ns	1.03
array/shared/copy	214958 ns	230791 ns	0.93
array/permutedims/4d	2471521 ns	2358000 ns	1.05
array/permutedims/2d	1170500 ns	1133208 ns	1.03
array/permutedims/3d	1772458 ns	1645604 ns	1.08
metal/synchronization/stream	19833.5 ns	18500 ns	1.07
metal/synchronization/context	20417 ns	19625 ns	1.04

This comment was automatically generated by workflow using github-action-benchmark.