Add arm64 to GHA CI and split tests across more runners #3158

andrjohns · 2025-03-05T03:05:01Z

Summary

Now that arm64 runners are available in github actions, we can add them as part of our regular CI. This will help catch issues like those fixed in #3059 - it's already identified issues with owens_t!

The owens_t function was hanging indefinitely under the arm64 runner, and this was resolved by adding our boost_policy definition as an argument - I'm not entirely sure why (might be another precision thing?).

I've added the runners as part of the existing Windows Rtools CI, using pwsh to avoid changing the current workflow in any meaningful way.

Tests

N/A - existing tests should still pass

Side Effects

N/A

Release notes

Add arm64 linux to Github Actions CI

Checklist

Copyright holder: Andrew Johnson

The copyright holder is typically you or your assignee, such as a university or company. By submitting this pull request, the copyright holder is agreeing to the license the submitted work under the following licenses:
- Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)
- Documentation: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
the basic tests are passing
- unit tests pass (to run, use: ./runTests.py test/unit)
- header checks pass, (make test-headers)
- dependencies checks pass, (make test-math-dependencies)
- docs build, (make doxygen)
- code passes the built in C++ standards checks (make cpplint)
the code is written in idiomatic C++ and changes are documented in the doxygen
the new changes are tested

stan-buildbot · 2025-03-05T09:04:07Z

Name	Old Result	New Result	Ratio	Performance change( 1 - new / old )
arma/arma.stan	0.36	0.42	0.87	-14.41% slower
low_dim_corr_gauss/low_dim_corr_gauss.stan	0.01	0.01	0.96	-3.72% slower
gp_regr/gen_gp_data.stan	0.03	0.02	1.06	5.94% faster
gp_regr/gp_regr.stan	0.1	0.1	1.02	2.4% faster
sir/sir.stan	73.23	71.53	1.02	2.32% faster
irt_2pl/irt_2pl.stan	4.29	4.34	0.99	-1.03% slower
eight_schools/eight_schools.stan	0.06	0.06	1.05	4.42% faster
pkpd/sim_one_comp_mm_elim_abs.stan	0.26	0.26	1.01	1.46% faster
pkpd/one_comp_mm_elim_abs.stan	20.25	20.02	1.01	1.14% faster
garch/garch.stan	0.44	0.43	1.0	0.46% faster
low_dim_gauss_mix/low_dim_gauss_mix.stan	2.84	2.77	1.03	2.52% faster
arK/arK.stan	1.85	1.81	1.02	2.04% faster
gp_pois_regr/gp_pois_regr.stan	3.04	2.82	1.08	7.31% faster
low_dim_gauss_mix_collapse/low_dim_gauss_mix_collapse.stan	8.9	8.58	1.04	3.61% faster
performance.compilation	193.67	191.38	1.01	1.19% faster
Mean result: 1.0127835456418288

Jenkins Console Log
Blue Ocean
Commit hash: 0f8dd4fc4b9973d67f612f68473daae100089e5c

Machine information

No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 20.04.3 LTS Release: 20.04 Codename: focal

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 46 bits physical, 48 bits virtual
CPU(s): 80
On-line CPU(s) list: 0-79
Thread(s) per core: 2
Core(s) per socket: 20
Socket(s): 2
NUMA node(s): 2
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
Stepping: 4
CPU MHz: 2400.000
CPU max MHz: 3700.0000
CPU min MHz: 1000.0000
BogoMIPS: 4800.00
Virtualization: VT-x
L1d cache: 1.3 MiB
L1i cache: 1.3 MiB
L2 cache: 40 MiB
L3 cache: 55 MiB
NUMA node0 CPU(s): 0,2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46,48,50,52,54,56,58,60,62,64,66,68,70,72,74,76,78
NUMA node1 CPU(s): 1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49,51,53,55,57,59,61,63,65,67,69,71,73,75,77,79
Vulnerability Gather data sampling: Mitigation; Microcode
Vulnerability Itlb multihit: KVM: Mitigation: VMX disabled
Vulnerability L1tf: Mitigation; PTE Inversion; VMX conditional cache flushes, SMT vulnerable
Vulnerability Mds: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Meltdown: Mitigation; PTI
Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Mitigation; IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; IBRS; IBPB conditional; STIBP conditional; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Mitigation; Clear CPU buffers; SMT vulnerable
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd mba ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req pku ospke md_clear flush_l1d arch_capabilities

G++:
g++ (Ubuntu 9.4.0-1ubuntu1~20.04) 9.4.0
Copyright (C) 2019 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Clang:
clang version 10.0.0-4ubuntu1
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin

andrjohns · 2025-03-05T14:49:47Z

I've also split the tests across more runners, bringing the total runtime down to ~1.5hours.

Given that the Jenkins mix unit tests take ~4 hours, maybe we could look at using the github runners for the basic unit tests instead (leaving the GPU, mpi, etc for Jenkins)? That could speed the CI process up a bit and also free up the resources for the CI in the other repos

WardBrian · 2025-03-05T14:52:04Z

We could also split up the Jenkins jobs into more parallel executors. Especially if they end up being short-lived, I don't think that would be a problem -- but I'd look into it in a separate PR rather than at the same time as this

WardBrian

Two minor comments but this looks good -- thanks!

.github/workflows/main.yml

stan-buildbot · 2025-03-06T01:07:06Z

Name	Old Result	New Result	Ratio	Performance change( 1 - new / old )
arma/arma.stan	0.38	0.42	0.91	-10.22% slower
low_dim_corr_gauss/low_dim_corr_gauss.stan	0.01	0.01	1.04	3.54% faster
gp_regr/gen_gp_data.stan	0.03	0.03	1.08	7.21% faster
gp_regr/gp_regr.stan	0.1	0.1	1.03	2.96% faster
sir/sir.stan	79.48	75.68	1.05	4.78% faster
irt_2pl/irt_2pl.stan	5.21	5.01	1.04	3.85% faster
eight_schools/eight_schools.stan	0.09	0.06	1.44	30.58% faster
pkpd/sim_one_comp_mm_elim_abs.stan	0.29	0.28	1.05	4.51% faster
pkpd/one_comp_mm_elim_abs.stan	21.51	21.85	0.98	-1.6% slower
garch/garch.stan	0.53	0.56	0.95	-4.95% slower
low_dim_gauss_mix/low_dim_gauss_mix.stan	2.96	3.09	0.96	-4.66% slower
arK/arK.stan	1.96	2.2	0.89	-11.97% slower
gp_pois_regr/gp_pois_regr.stan	3.09	3.45	0.9	-11.56% slower
low_dim_gauss_mix_collapse/low_dim_gauss_mix_collapse.stan	9.73	9.38	1.04	3.58% faster
performance.compilation	203.88	205.61	0.99	-0.85% slower
Mean result: 1.0227237936926816

Jenkins Console Log
Blue Ocean
Commit hash: 41fb879c2b057722f0b8d4e3462ddede193cb95e

Machine information

No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 20.04.3 LTS Release: 20.04 Codename: focal

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 46 bits physical, 48 bits virtual
CPU(s): 80
On-line CPU(s) list: 0-79
Thread(s) per core: 2
Core(s) per socket: 20
Socket(s): 2
NUMA node(s): 2
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
Stepping: 4
CPU MHz: 2400.000
CPU max MHz: 3700.0000
CPU min MHz: 1000.0000
BogoMIPS: 4800.00
Virtualization: VT-x
L1d cache: 1.3 MiB
L1i cache: 1.3 MiB
L2 cache: 40 MiB
L3 cache: 55 MiB
NUMA node0 CPU(s): 0,2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46,48,50,52,54,56,58,60,62,64,66,68,70,72,74,76,78
NUMA node1 CPU(s): 1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49,51,53,55,57,59,61,63,65,67,69,71,73,75,77,79
Vulnerability Gather data sampling: Mitigation; Microcode
Vulnerability Itlb multihit: KVM: Mitigation: VMX disabled
Vulnerability L1tf: Mitigation; PTE Inversion; VMX conditional cache flushes, SMT vulnerable
Vulnerability Mds: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Meltdown: Mitigation; PTI
Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Mitigation; IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; IBRS; IBPB conditional; STIBP conditional; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Mitigation; Clear CPU buffers; SMT vulnerable
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd mba ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req pku ospke md_clear flush_l1d arch_capabilities

G++:
g++ (Ubuntu 9.4.0-1ubuntu1~20.04) 9.4.0
Copyright (C) 2019 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Clang:
clang version 10.0.0-4ubuntu1
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin

stan-buildbot · 2025-03-06T21:11:47Z

Name	Old Result	New Result	Ratio	Performance change( 1 - new / old )
arma/arma.stan	0.36	0.34	1.07	6.32% faster
low_dim_corr_gauss/low_dim_corr_gauss.stan	0.01	0.01	0.95	-5.19% slower
gp_regr/gen_gp_data.stan	0.03	0.03	1.02	2.07% faster
gp_regr/gp_regr.stan	0.1	0.1	1.03	3.25% faster
sir/sir.stan	74.1	70.84	1.05	4.4% faster
irt_2pl/irt_2pl.stan	4.22	4.3	0.98	-1.78% slower
eight_schools/eight_schools.stan	0.06	0.06	0.99	-1.24% slower
pkpd/sim_one_comp_mm_elim_abs.stan	0.26	0.27	0.97	-3.11% slower
pkpd/one_comp_mm_elim_abs.stan	19.76	19.99	0.99	-1.17% slower
garch/garch.stan	0.44	0.43	1.01	1.47% faster
low_dim_gauss_mix/low_dim_gauss_mix.stan	2.75	2.73	1.01	0.77% faster
arK/arK.stan	1.83	1.86	0.99	-1.4% slower
gp_pois_regr/gp_pois_regr.stan	2.91	2.92	1.0	-0.2% slower
low_dim_gauss_mix_collapse/low_dim_gauss_mix_collapse.stan	8.92	8.85	1.01	0.8% faster
performance.compilation	193.18	193.84	1.0	-0.34% slower
Mean result: 1.0039277839493728

Jenkins Console Log
Blue Ocean
Commit hash: 41fb879c2b057722f0b8d4e3462ddede193cb95e

Machine information

No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 20.04.3 LTS Release: 20.04 Codename: focal

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 46 bits physical, 48 bits virtual
CPU(s): 80
On-line CPU(s) list: 0-79
Thread(s) per core: 2
Core(s) per socket: 20
Socket(s): 2
NUMA node(s): 2
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
Stepping: 4
CPU MHz: 2400.000
CPU max MHz: 3700.0000
CPU min MHz: 1000.0000
BogoMIPS: 4800.00
Virtualization: VT-x
L1d cache: 1.3 MiB
L1i cache: 1.3 MiB
L2 cache: 40 MiB
L3 cache: 55 MiB
NUMA node0 CPU(s): 0,2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46,48,50,52,54,56,58,60,62,64,66,68,70,72,74,76,78
NUMA node1 CPU(s): 1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49,51,53,55,57,59,61,63,65,67,69,71,73,75,77,79
Vulnerability Gather data sampling: Mitigation; Microcode
Vulnerability Itlb multihit: KVM: Mitigation: VMX disabled
Vulnerability L1tf: Mitigation; PTE Inversion; VMX conditional cache flushes, SMT vulnerable
Vulnerability Mds: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Meltdown: Mitigation; PTI
Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Mitigation; IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; IBRS; IBPB conditional; STIBP conditional; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Mitigation; Clear CPU buffers; SMT vulnerable
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd mba ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req pku ospke md_clear flush_l1d arch_capabilities

G++:
g++ (Ubuntu 9.4.0-1ubuntu1~20.04) 9.4.0
Copyright (C) 2019 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Clang:
clang version 10.0.0-4ubuntu1
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin

WardBrian · 2025-03-07T21:34:44Z

Hey @andrjohns, I've noticed some extra flakiness in the windows tests after this -- e.g,
https://github.com/stan-dev/math/actions/runs/13723040759/attempts/1
and the re-run:
https://github.com/stan-dev/math/actions/runs/13723040759/job/38398633425 (notice it's a different test that fails here)

These passed during the CI of #3151, which was using the old action at the time. Any idea what's up?

WardBrian · 2025-03-08T01:32:55Z

Update: Run #4 did pass. Definitely worth looking into more next week

andrjohns added 2 commits March 5, 2025 08:38

Add arm64 CI, fix owens_t hang

a03c0e3

Backticks for pwsh

9521ef2

andrjohns added 6 commits March 5, 2025 17:26

Simplify file and expand parallelism

d51d38c

Update labels

3ef4c1a

Update mix/fun parallelism

478da34

Update mix parallelism

7a4ea21

Don't fail fast

134e6b7

Further split mix tests

117b951

andrjohns changed the title ~~Add arm64 to the GHA CI~~ Add arm64 to GHA CI and split tests across more runners Mar 5, 2025

WardBrian reviewed Mar 5, 2025

View reviewed changes

.github/workflows/main.yml Outdated Show resolved Hide resolved

.github/workflows/main.yml Show resolved Hide resolved

WardBrian mentioned this pull request Mar 5, 2025

Segfaulting compiler in github CI on arm roualdes/bridgestan#268

Closed

Simplify matrix config, update comments

c902224

WardBrian approved these changes Mar 6, 2025

View reviewed changes

WardBrian merged commit 5dc041b into develop Mar 6, 2025
24 checks passed

WardBrian deleted the arm64-test branch March 6, 2025 21:20

WardBrian mentioned this pull request Mar 27, 2025

Floating point precision loss under arm64 (FP_CONTRACT) #2935

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Add arm64 to GHA CI and split tests across more runners #3158

Add arm64 to GHA CI and split tests across more runners #3158

Uh oh!

andrjohns commented Mar 5, 2025

Uh oh!

stan-buildbot commented Mar 5, 2025

Uh oh!

andrjohns commented Mar 5, 2025

Uh oh!

WardBrian commented Mar 5, 2025

Uh oh!

WardBrian left a comment

Uh oh!

Uh oh!

Uh oh!

stan-buildbot commented Mar 6, 2025

Uh oh!

stan-buildbot commented Mar 6, 2025

Uh oh!

Uh oh!

WardBrian commented Mar 7, 2025

Uh oh!

WardBrian commented Mar 8, 2025

Uh oh!

Uh oh!

Uh oh!

Add arm64 to GHA CI and split tests across more runners #3158

Add arm64 to GHA CI and split tests across more runners #3158

Uh oh!

Conversation

andrjohns commented Mar 5, 2025

Summary

Tests

Side Effects

Release notes

Checklist

Uh oh!

stan-buildbot commented Mar 5, 2025

Uh oh!

andrjohns commented Mar 5, 2025

Uh oh!

WardBrian commented Mar 5, 2025

Uh oh!

WardBrian left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

stan-buildbot commented Mar 6, 2025

Uh oh!

stan-buildbot commented Mar 6, 2025

Uh oh!

Uh oh!

WardBrian commented Mar 7, 2025

Uh oh!

WardBrian commented Mar 8, 2025

Uh oh!

Uh oh!