Skip to content

Update asv #2437

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jeromekelleher merged 1 commit into from
Dec 9, 2025
Merged

Update asv #2437

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
275 changes: 134 additions & 141 deletions benchmarks/benchmarks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,6 +24,7 @@
documentation for details on how to run these and how to develop your
own benchmarks.
"""

try:
import stdpopsim

Expand Down Expand Up @@ -55,148 +56,140 @@ def setup(self):
self.recomb_map_chr22 = genetic_map.get_chromosome_map("chr22")


class Hudson(LargeSimulationBenchmark):
def _run_large_sample_size(self):
msprime.simulate(
sample_size=10**6,
length=1e7,
Ne=10**4,
recombination_rate=1e-8,
random_seed=42,
)

def time_large_sample_size(self):
self._run_large_sample_size()

def peakmem_large_sample_size(self):
self._run_large_sample_size()

def _run_long_sequence_length(self):
msprime.simulate(
sample_size=100,
length=1e8,
Ne=10**4,
recombination_rate=1e-8,
random_seed=42,
)

def time_long_sequence_length(self):
self._run_long_sequence_length()

def peakmem_long_sequence_length(self):
self._run_long_sequence_length()

def _run_long_sequence_length_gene_conversion(self):
msprime.sim_ancestry(
sample_size=100,
length=1e8,
Ne=10**4,
gene_conversion_rate=1e-8,
# 100Kb tract length.
gene_conversion_tract_length=100 * 1e3,
random_seed=43,
)

def time_long_sequence_length_gene_conversion(self):
self._run_long_sequence_length()

def peakmem_long_sequence_length_gene_conversion(self):
self._run_long_sequence_length()

def _run_human_chr22(self):
msprime.simulate(
sample_size=100,
Ne=10**4,
recombination_map=self.recomb_map_chr22,
random_seed=234,
)

def time_human_chr22(self):
self._run_human_chr22()

def peakmem_human_chr22(self):
self._run_human_chr22()

def _run_many_replicates(self):
for _ in msprime.simulate(10, num_replicates=10**5, random_seed=1234):
pass
class HudsonlargeSampleSize(LargeSimulationBenchmark):

def _get_params(self):
return {
"samples": 0.5 * (10**6),
"sequence_length": 1e7,
"population_size": 10**4,
"recombination_rate": 1e-8,
"random_seed": 42,
}

def run(self):
return msprime.sim_ancestry(**self._get_params())

def time_test(self):
self.run()

def peakmem_test(self):
self.run()


class HudsonlargeSampleSizeOverRoot(HudsonlargeSampleSize):
def _get_params(self):
return {
**super()._get_params(),
"stop_at_local_mrca": False,
}


def time_many_replicates(self):
self._run_many_replicates()

def peakmem_many_replicates(self):
self._run_many_replicates()

# 2 populations, high migration.
# Lots of populations, 1D stepping stone.


class DTWF(LargeSimulationBenchmark):
def _run_large_population_size(self):
msprime.simulate(
sample_size=1000,
Ne=10**6,
length=1e5,
recombination_rate=1e-8,
random_seed=42,
model="dtwf",
end_time=1000,
)

def time_large_population_size(self):
self._run_large_population_size()

def peakmem_large_population_size(self):
self._run_large_population_size()

def _run_long_sequence_length(self):
msprime.simulate(
sample_size=100,
Ne=10**4,
length=1e7,
recombination_rate=1e-8,
random_seed=42,
model="dtwf",
# Tuning this to give ~30s runtime.
end_time=5e4,
)

def time_long_sequence_length(self):
self._run_long_sequence_length()

def peakmem_long_sequence_length(self):
self._run_long_sequence_length()

def _run_human_chr22(self):
msprime.simulate(
sample_size=100,
Ne=10**4,
recombination_map=self.recomb_map_chr22,
random_seed=234,
end_time=10000,
model="dtwf",
)

def time_human_chr22(self):
self._run_human_chr22()

def peakmem_human_chr22(self):
self._run_human_chr22()

def _run_many_replicates(self):
reps = msprime.simulate(
10,
Ne=100,
num_replicates=10**5,
random_seed=1234,
model="dtwf",
end_time=100,
)
for _ in reps:
class HudsonLongSequenceLength(HudsonlargeSampleSize):
def _get_params(self):
return {
**super()._get_params(),
"sequence_length": 1e8,
"samples": 50,
}


class HudsonLongSequenceLengthGeneConversion(HudsonlargeSampleSize):
def _get_params(self):
return {
"sequence_length": 1e8,
"samples": 50,
"gene_conversion_rate": 1e-8,
"gene_conversion_tract_length": 100 * 1e3,
"random_seed": 43,
}


class HudsonHumanChr22(HudsonlargeSampleSize):

def _get_params(self):
return {
**super()._get_params(),
"sequence_length": None,
"samples": 50,
"recombination_rate": self.recomb_map_chr22,
}


class HudsonManyReplicates(HudsonlargeSampleSize):

def run(self):
params = {"samples": 10, "num_replicates": 10**5, "random_seed": 1234}
for _ in msprime.sim_ancestry(**params):
pass

def time_many_replicates(self):
self._run_many_replicates()

def peakmem_many_replicates(self):
self._run_many_replicates()
class HudsonHumanChr22OverRoot(HudsonlargeSampleSize):
def _get_params(self):
return {
**super()._get_params(),
"sequence_length": None,
"samples": 50,
"recombination_rate": self.recomb_map_chr22,
"stop_at_local_mrca": False,
}


class DTWFLargePopulationSize(LargeSimulationBenchmark):
def _get_params(self):
return {
"samples": 500,
"sequence_length": 1e5,
"population_size": 10**6,
"recombination_rate": 1e-8,
"random_seed": 42,
"model": "dtwf",
"end_time": 1000,
}

def run(self):
return msprime.sim_ancestry(**self._get_params())

def time_test(self):
self.run()

def peakmem_test(self):
self.run()


class DTWFLongSequenceLength(DTWFLargePopulationSize):
def _get_params(self):
return {
**super()._get_params(),
"sequence_length": 1e7,
"samples": 50,
"end_time": 5e4,
"population_size": 10**4,
}


class DTWFHumanChr22(DTWFLargePopulationSize):

def _get_params(self):
return {
**super()._get_params(),
"sequence_length": None,
"samples": 50,
"recombination_rate": self.recomb_map_chr22,
"end_time": 10000,
"population_size": 10**4,
}


class DTWFManyReplicates(DTWFLargePopulationSize):
def run(self):
params = {
"samples": 5,
"population_size": 100,
"num_replicates": 10**5,
"random_seed": 1234,
"model": "dtwf",
"end_time": 100,
}
for _ in msprime.sim_ancestry(**params):
pass
4 changes: 4 additions & 0 deletions benchmarks/check_asv.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,10 @@ sudo cpufreq-set -c 3 -g performance
#asv does recent commits first, so by letting it run for 55min, and the cron to 1hr
#it will always keep up with new commits, but also process the backlog
timeout 3300s asv run -j 4 --show-stderr --cpu-affinity 3 --skip-existing ALL

#instead you can test for a set of versions: uncomment the next line and comment line 6 to do so
#you can add more version to benshmark to the file version_hashes.txt
#asv run -j 4 --show-stderr --cpu-affinity 3 --skip-existing HASHFILE:benchmarks/version_hashes.txt
sudo cpufreq-set -c 3 -g powersave
asv publish
asv preview
16 changes: 10 additions & 6 deletions benchmarks/version_hashes.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,10 @@
661200e8fc310d374eb266cad39d4d3d96f0b8e9
8ce627f754ee52d7961fd9016fcbed202234734e
8a39c401d1c817f4950b7fc486da8238b9fa76ec
3e28f9dea511e6745f99cbf2c009cdcea4a73f20
b8f5fb60b0ca0885b7f1c2d7ac7a9ae3b95fc543
18d3c417db2cdd87ad6539a5e2cef60250460b3a
d2e99ab2d4ba828bc1f19b0eb23f0334c1b26f1a
cda89ed3f670c428d0452607753dccd8fc79f5ff
2ddc65cc1e83c6d3c2af78e781bd5983a0364ff7
0da0150a2d65053933c2c5f9b598e9197c7ea9ac
ad3f625715821a3133d841a70d2a782d7e8fd171
ff3f7851b647f28640cc2cbca62cfd056508e2d6
804e0361c4f8b5f5051a9fbf411054ee8be3426a
c11a2f12c72dd054a0ade0767474184ceec8281c
f6fc608975bce01070f23e785d89fe171d190a9a
57ef4ee3267cd9b8e711787539007b0cde94c55c