Fix race conditions caused by missing fences in OpenCL kernels for cholesky_decompose, tridiagonalization and diag_inv

[jachymb]

Introduction

On my machine a noticed that the GPU kernel for cholesky_decompose is buggy and actually produces wrong results. I confirmed with Oclgrind that this is due to a fencing race condition. The attached test fails on my Radeon GPU against current develop and is fixed with this patch. I then tested other kernels in the repo with Oclgrind and found possible fencing race conditions in the kernels for tridiagonalization and diag_inv. For these two, I was unable to construct a failing test so maybe they are just latent issues. However, this kind of problems may be machine-specific, so I believe it's better fixed.

This fixes #3429. See also #3430.

The rest of this commentary was made by Fable5.1 which helped me debug this.

Summary

In three kernels, work items read or overwrite buffer entries that other work items access on the other side of a barrier(CLK_LOCAL_MEM_FENCE). That flag orders operations on local memory only; such accesses need CLK_GLOBAL_MEM_FENCE.

• cholesky_decompose: both barriers now fence global memory. This kernel returns wrong results on an AMD GPU. With more than 64 work items, some of them read entries of the matrix as they were before other work items overwrote them, so cholesky_decompose of a matrix_cl with more than 64 rows silently returns a wrong factor.
• tridiagonalization_householder: the barrier after the first loop now fences local and global memory.
• diag_inv: a barrier(CLK_GLOBAL_MEM_FENCE) now separates the reads of the matrix from the final copy into it.

The last two were found with the race detector of Oclgrind (oclgrind --data-races <test>); no wrong result from them has been seen. On the small cases of their tests, Oclgrind reports 14, 62 and more than 1000 races in the three kernels on develop and none with this PR.

Tests

• New: MathMatrixOpenCL.cholesky_decompose_cpu_vs_cl_one_work_group factors a 200Ă—200 matrix, which is one call of the kernel with 200 work items. On the AMD GPU it fails on develop in 5 of 5 runs and passes with the fix.
• No new test for the other two kernels: no device at hand returns a wrong result from them, so a test could not fail on develop. The Oclgrind reports are the evidence.
• The existing OpenCL tests of the Cholesky decomposition, the triangular inverse and the tridiagonalization pass on two CPU OpenCL devices (PoCL, Intel) and on the AMD GPU.

Side Effects

• Results are identical to develop wherever develop was correct.
• Speed: no measurable change on the CPU OpenCL devices. On the AMD GPU the Cholesky kernel is 2 to 5 % slower for blocks of 128 to 256 rows and up to 9 % for 64 rows; the triangular inverse and the tridiagonalization show no change beyond the scatter between runs.

Release notes

Fix race conditions caused by missing fences in OpenCL kernels for cholesky_decompose, tridiagonalization and diag_inv

Checklist

• Copyright holder: Jáchym Barvínek jachymb@gmail.com

  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]

Name	Old Result	New Result	Ratio	Performance change( 1 - new / old )
stat_comp_benchmarks/benchmarks/gp_regr/gp_regr.stan	0.23	0.23	0.99	-1.21% slower
stat_comp_benchmarks/benchmarks/gp_regr/gen_gp_data.stan	0.06	0.06	0.99	-0.77% slower
stat_comp_benchmarks/benchmarks/low_dim_corr_gauss/low_dim_corr_gauss.stan	0.02	0.02	1.0	-0.19% slower
stat_comp_benchmarks/benchmarks/arma/arma.stan	0.71	0.71	1.0	0.31% faster
stat_comp_benchmarks/benchmarks/arK/arK.stan	3.2	3.19	1.0	0.21% faster
stat_comp_benchmarks/benchmarks/pkpd/one_comp_mm_elim_abs.stan	42.64	42.7	1.0	-0.12% slower
stat_comp_benchmarks/benchmarks/pkpd/sim_one_comp_mm_elim_abs.stan	0.6	0.6	1.0	-0.23% slower
stat_comp_benchmarks/benchmarks/gp_pois_regr/gp_pois_regr.stan	4.48	4.46	1.0	0.44% faster
stat_comp_benchmarks/benchmarks/low_dim_gauss_mix_collapse/low_dim_gauss_mix_collapse.stan	21.03	20.86	1.01	0.8% faster
stat_comp_benchmarks/benchmarks/garch/garch.stan	0.89	0.89	1.0	-0.22% slower
stat_comp_benchmarks/benchmarks/eight_schools/eight_schools.stan	0.11	0.11	1.01	1.48% faster
stat_comp_benchmarks/benchmarks/irt_2pl/irt_2pl.stan	8.45	8.45	1.0	0.05% faster
stat_comp_benchmarks/benchmarks/low_dim_gauss_mix/low_dim_gauss_mix.stan	6.35	6.32	1.01	0.54% faster
stat_comp_benchmarks/benchmarks/sir/sir.stan	165.58	165.78	1.0	-0.12% slower
performance.compilation	379.51	385.46	0.98	-1.57% slower
Mean result: 0.9996594960254679

Jenkins Console Log
Jenkins Build Stages
Commit hash: 993f35f70b6d7c411c87d88bbb003259bfff0fcd

Machine information

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:                           43 bits physical, 48 bits virtual
CPU(s):                                  256
On-line CPU(s) list:                     0-255
Thread(s) per core:                      2
Core(s) per socket:                      64
Socket(s):                               2
NUMA node(s):                            2
Vendor ID:                               AuthenticAMD
CPU family:                              23
Model:                                   49
Model name:                              AMD EPYC 7742 64-Core Processor
Stepping:                                0
Frequency boost:                         enabled
CPU MHz:                                 1497.090
CPU max MHz:                             3416.0681
CPU min MHz:                             1500.0000
BogoMIPS:                                4491.56
Virtualization:                          AMD-V
L1d cache:                               4 MiB
L1i cache:                               4 MiB
L2 cache:                                64 MiB
L3 cache:                                512 MiB
NUMA node0 CPU(s):                       0-63,128-191
NUMA node1 CPU(s):                       64-127,192-255
Vulnerability Gather data sampling:      Not affected
Vulnerability Indirect target selection: Not affected
Vulnerability Itlb multihit:             Not affected
Vulnerability L1tf:                      Not affected
Vulnerability Mds:                       Not affected
Vulnerability Meltdown:                  Not affected
Vulnerability Mmio stale data:           Not affected
Vulnerability Old microcode:             Not affected
Vulnerability Reg file data sampling:    Not affected
Vulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection
Vulnerability Spec rstack overflow:      Mitigation; Safe RET
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; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds:                     Not affected
Vulnerability Tsa:                       Not affected
Vulnerability Tsx async abort:           Not affected
Vulnerability Vmscape:                   Mitigation; IBPB before exit to userspace
Flags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es

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