{perf}[gompi/2025b] HPCToolkit v2025.0.1 w/ CUDA 12.9.1 by Thyre · Pull Request #23830 · easybuilders/easybuild-easyconfigs

Thyre · 2025-09-09T12:42:56Z

(created using eb --new-pr)

Requires:

TODO:

Test on aarch64
Enable tests and run them
Check if we can move some of the dependencies to builddependencies

HPCViewer, while also interesting to have, needs to wait until we have GTK3/GTK4. This is blocked by a few more missing dependencies for now.

…25.0.0-gompi-2025b-CUDA-12.9.1.eb

github-actions · 2025-09-09T12:43:20Z

Updated software `HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb`

Diff against HPCToolkit-2025.0.1-gompi-2025b.eb

easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b.eb

diff --git a/easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b.eb b/easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb
index ce71f009ef..46a7b253f5 100644
--- a/easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b.eb
+++ b/easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb
@@ -2,6 +2,7 @@ easyblock = 'MesonNinja'
 
 name = 'HPCToolkit'
 version = '2025.0.1'
+versionsuffix = '-CUDA-%(cudaver)s'
 
 homepage = 'https://hpctoolkit.org/'
 description = """
@@ -41,6 +42,7 @@ builddependencies = [
 
 dependencies = [
     ('Boost', '1.88.0'),
+    ('CUDA', '12.9.1', '', SYSTEM),
     ('Dyninst', '13.0.0'),
     ('PAPI', '7.2.0'),
     ('Xerces-C++', '3.3.0'),
@@ -54,7 +56,7 @@ if ARCH == 'x86_64':
     dependencies.append(('intel-XED', '2025.06.08'))
 
 configopts = "-Dtests=enabled -Dmanpages=disabled -Dmanual=disabled -Dhpcprof_mpi=enabled "
-configopts += "-Dpapi=enabled -Dcuda=disabled -Dgtpin=disabled -Drocm=disabled -Dopencl=disabled "
+configopts += "-Dpapi=enabled -Dcuda=enabled -Dgtpin=disabled -Drocm=disabled -Dopencl=disabled "
 configopts += "-Dvalgrind_annotations=false -Dpython=disabled "
 
 # Explicitly ensure that no environment variables related to jobs are picked up, as this
@@ -71,6 +73,12 @@ local_env_to_unset = [
 pretestopts = ""
 for local_env in local_env_to_unset:
     pretestopts += f"unset {local_env} && "
+
+build_info_msg = """CUDA related tests can fail due to insufficient permissions for CUPTI PC Sampling.
+The final installation is still usable, but with reduced functionality.
+See: https://hpctoolkit.gitlab.io/hpctoolkit/2025.0.1/users/faq.html
+in section 'Ensuring permission to use GPU performance counters'
+Building with '--ignore-test-failure' still results in a working installation"""
 runtest = "meson test"
 
 local_preload_libs = [
@@ -81,11 +89,10 @@ sanity_check_paths = {
     'files':
         ['bin/hpcrun', 'include/hpctoolkit.h'] +
         ['lib/libhpcrun_preload_%s.%s' % (a, SHLIB_EXT) for a in local_preload_libs] +
-        [f'lib/libhpcrun.{SHLIB_EXT}', f'lib/libhpctoolkit.{SHLIB_EXT}'],
+        [f'lib/libhpcrun.{SHLIB_EXT}', f'lib/libhpctoolkit.{SHLIB_EXT}', f'lib/libhpcrun_dlopen_nvidia.{SHLIB_EXT}'],
     'dirs': [],
 }
 
-# hpcstructs version command exits with 1, so grep for the version instead
 sanity_check_commands = [
     'hpcrun --version',
     'hpcstruct --version | grep %(version)s',

Updated software `HPCToolkit-2025.0.1-gompi-2025b.eb`

Diff against HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb

easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb

diff --git a/easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb b/easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b.eb
index 46a7b253f5..ce71f009ef 100644
--- a/easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb
+++ b/easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b.eb
@@ -2,7 +2,6 @@ easyblock = 'MesonNinja'
 
 name = 'HPCToolkit'
 version = '2025.0.1'
-versionsuffix = '-CUDA-%(cudaver)s'
 
 homepage = 'https://hpctoolkit.org/'
 description = """
@@ -42,7 +41,6 @@ builddependencies = [
 
 dependencies = [
     ('Boost', '1.88.0'),
-    ('CUDA', '12.9.1', '', SYSTEM),
     ('Dyninst', '13.0.0'),
     ('PAPI', '7.2.0'),
     ('Xerces-C++', '3.3.0'),
@@ -56,7 +54,7 @@ if ARCH == 'x86_64':
     dependencies.append(('intel-XED', '2025.06.08'))
 
 configopts = "-Dtests=enabled -Dmanpages=disabled -Dmanual=disabled -Dhpcprof_mpi=enabled "
-configopts += "-Dpapi=enabled -Dcuda=enabled -Dgtpin=disabled -Drocm=disabled -Dopencl=disabled "
+configopts += "-Dpapi=enabled -Dcuda=disabled -Dgtpin=disabled -Drocm=disabled -Dopencl=disabled "
 configopts += "-Dvalgrind_annotations=false -Dpython=disabled "
 
 # Explicitly ensure that no environment variables related to jobs are picked up, as this
@@ -73,12 +71,6 @@ local_env_to_unset = [
 pretestopts = ""
 for local_env in local_env_to_unset:
     pretestopts += f"unset {local_env} && "
-
-build_info_msg = """CUDA related tests can fail due to insufficient permissions for CUPTI PC Sampling.
-The final installation is still usable, but with reduced functionality.
-See: https://hpctoolkit.gitlab.io/hpctoolkit/2025.0.1/users/faq.html
-in section 'Ensuring permission to use GPU performance counters'
-Building with '--ignore-test-failure' still results in a working installation"""
 runtest = "meson test"
 
 local_preload_libs = [
@@ -89,10 +81,11 @@ sanity_check_paths = {
     'files':
         ['bin/hpcrun', 'include/hpctoolkit.h'] +
         ['lib/libhpcrun_preload_%s.%s' % (a, SHLIB_EXT) for a in local_preload_libs] +
-        [f'lib/libhpcrun.{SHLIB_EXT}', f'lib/libhpctoolkit.{SHLIB_EXT}', f'lib/libhpcrun_dlopen_nvidia.{SHLIB_EXT}'],
+        [f'lib/libhpcrun.{SHLIB_EXT}', f'lib/libhpctoolkit.{SHLIB_EXT}'],
     'dirs': [],
 }
 
+# hpcstructs version command exits with 1, so grep for the version instead
 sanity_check_commands = [
     'hpcrun --version',
     'hpcstruct --version | grep %(version)s',

Fail tests due to incorrect values Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

May pick up system OpenCL headers. Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Thyre · 2025-09-09T13:51:30Z

Test report by @Thyre
SUCCESS
Build succeeded for 2 out of 2 (2 easyconfigs in total)
ZAM054 - Linux Zorin OS 17, x86_64, 12th Gen Intel(R) Core(TM) i7-1260P, 1 x NVIDIA NVIDIA GeForce MX550, 580.65.06, Python 3.10.12
See https://gist.github.com/Thyre/4de99d300e14b77888b5cacbbc8cbf50 for a full test report.

Thyre · 2025-09-09T13:57:11Z

hpctoolkit 2025.0.0

  Subprojects
    hpctesttool-wheels  : YES
    md5-g72cf2cd        : YES
    valgrind-headers    : YES

  User defined options
    b_ndebug            : true
    cuda                : enabled
    gtpin               : disabled
    hpcprof_mpi         : enabled
    libdir              : lib
    manpages            : disabled
    manual              : disabled
    opencl              : disabled
    optimization        : 2
    papi                : enabled
    prefix              : /opt/EasyBuild/apps/software/HPCToolkit/2025.0.0-gompi-2025b-CUDA-12.9.1
    python              : disabled
    rocm                : disabled
    tests               : enabled
    valgrind_annotations: false

The three mentioned subprojects are all part of the HPCToolkit sources. So it shouldn't try to download anything fortunately.

Thyre · 2025-09-09T14:01:23Z

Test report by @Thyre
FAILED
Build succeeded for 1 out of 3 (2 easyconfigs in total)
jrc0900.jureca - Linux Rocky Linux 9.6, AArch64, ARM UNKNOWN, 1 x NVIDIA NVIDIA GH200 480GB, 580.65.06, Python 3.9.21
See https://gist.github.com/Thyre/8f816e53070a7ffd6aac94a191ed93cb for a full test report.

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Thyre · 2025-09-09T14:33:04Z

Test report by @Thyre
FAILED
Build succeeded for 1 out of 3 (2 easyconfigs in total)
jrc0900.jureca - Linux Rocky Linux 9.6, AArch64, ARM UNKNOWN, 1 x NVIDIA NVIDIA GH200 480GB, 580.65.06, Python 3.9.21
See https://gist.github.com/Thyre/8f816e53070a7ffd6aac94a191ed93cb for a full test report.

Tests here failed because I was building on a node, which was picked up and is not correctly handled by the check.
I also tried on a login node of JUPITER. Here, tests failed because of insufficient permissions for PCSampling.
I would propose to keep the tests enabled, but add a comment noting which failures one can expect to see.

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

…ew_pr_HPCToolkit202500

Thyre · 2025-09-09T14:36:00Z

@boegelbot please test @ jsc-zen3-a100

boegelbot · 2025-09-09T14:40:09Z

@Thyre: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de

PR test command 'if [[ develop != 'develop' ]]; then EB_BRANCH=develop ./easybuild_develop.sh 2> /dev/null 1>&2; EB_PREFIX=/home/boegelbot/easybuild/develop source init_env_easybuild_develop.sh; fi; EB_PR=23830 EB_ARGS= EB_CONTAINER= EB_REPO=easybuild-easyconfigs EB_BRANCH=develop /opt/software/slurm/bin/sbatch --job-name test_PR_23830 --ntasks=8 --partition=jsczen3g --gres=gpu:1 ~/boegelbot/eb_from_pr_upload_jsc-zen3.sh' executed!

exit code: 0
output:

Submitted batch job 7898

Test results coming soon (I hope)...

Details

- notification for comment with ID 3271028326 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

Thyre · 2025-09-09T14:40:57Z

HPCToolkit-2025.0.0-gompi-2025b.eb

Test report by @Thyre
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
jpbl-s01-04 - Linux RHEL 9.5, AArch64, ARM UNKNOWN (neoverse_v2), 1 x NVIDIA NVIDIA GH200 480GB, 570.133.20, Python 3.9.21
See https://gist.github.com/Thyre/acd8bb52177cbf4e51e17fc5ee2e6e54 for a full test report.

Thyre · 2025-09-09T14:42:18Z

Test report by @Thyre
SUCCESS
Build succeeded for 2 out of 2 (2 easyconfigs in total)
ZAM054 - Linux Zorin OS 17, x86_64, 12th Gen Intel(R) Core(TM) i7-1260P, 1 x NVIDIA NVIDIA GeForce MX550, 580.65.06, Python 3.10.12
See https://gist.github.com/Thyre/e21462e1cd3d43e946a8eb033cb00b7d for a full test report.

Thyre · 2025-09-09T14:45:20Z

HPCToolkit-2025.0.0-gompi-2025b-CUDA-12.9.1.eb
Missing permissions for CUPTI PC Sampling

Test report by @Thyre
FAILED
Build succeeded for 0 out of 1 (1 easyconfigs in total)
jpbl-s01-04 - Linux RHEL 9.5, AArch64, ARM UNKNOWN (neoverse_v2), 1 x NVIDIA NVIDIA GH200 480GB, 570.133.20, Python 3.9.21
See https://gist.github.com/Thyre/f5d7ebda39b0ea68f1502a94613a1fbf for a full test report.

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

boegelbot · 2025-09-09T14:49:12Z

Test report by @boegelbot
FAILED
Build succeeded for 0 out of 2 (2 easyconfigs in total)
jsczen3g1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.6, x86_64, AMD EPYC-Milan Processor (zen3), 1 x NVIDIA NVIDIA A100 80GB PCIe, 575.57.08, Python 3.9.21
See https://gist.github.com/boegelbot/7e4342c56c227bf4df6e966833e9bbbb for a full test report.

Thyre · 2025-09-09T14:50:13Z

Test report by boegelbot
FAILED
Build succeeded for 0 out of 2 (2 easyconfigs in total)
jsczen3g1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.6, x86_64, AMD EPYC-Milan Processor (zen3), 1 x NVIDIA NVIDIA A100 80GB PCIe, 575.57.08, Python 3.9.21
See https://gist.github.com/boegelbot/7e4342c56c227bf4df6e966833e9bbbb for a full test report.

Yeah... that's HPCToolkit picking up that we're running on a compute node. The CUDA tests also fail with insufficient permissions.

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Thyre · 2025-09-10T12:52:25Z

Test report by @Thyre
FAILED
Build succeeded for 0 out of 2 (2 easyconfigs in total)
jrc0900.jureca - Linux Rocky Linux 9.6, AArch64, ARM UNKNOWN, 1 x NVIDIA NVIDIA GH200 480GB, 580.65.06, Python 3.9.21
See https://gist.github.com/Thyre/14366173be5b4a5eee08bf7a96c13d65 for a full test report.

stripped See https://gitlab.com/hpctoolkit/hpctoolkit/-/merge_requests/1310 for more information on why this is done. Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Thyre · 2025-09-10T13:34:46Z

Test report by @Thyre
SUCCESS
Build succeeded for 2 out of 2 (2 easyconfigs in total)
ZAM054 - Linux Zorin OS 17, x86_64, 12th Gen Intel(R) Core(TM) i7-1260P, 1 x NVIDIA NVIDIA GeForce MX550, 580.65.06, Python 3.10.12
See https://gist.github.com/Thyre/7bc822c132f5f3581fdfd88afa6a1c8f for a full test report.

Thyre · 2025-09-10T13:36:43Z

Test report by @Thyre
SUCCESS
Build succeeded for 2 out of 2 (2 easyconfigs in total)
jrc0900.jureca - Linux Rocky Linux 9.6, AArch64, ARM UNKNOWN, 1 x NVIDIA NVIDIA GH200 480GB, 580.65.06, Python 3.9.21
See https://gist.github.com/Thyre/ce73e0ce1c6578b00795ec7657a63d75 for a full test report.

Thyre · 2025-09-10T13:37:28Z

@boegelbot please test @ jsc-zen3-a100

boegelbot · 2025-09-10T13:40:11Z

@Thyre: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de

PR test command 'if [[ develop != 'develop' ]]; then EB_BRANCH=develop ./easybuild_develop.sh 2> /dev/null 1>&2; EB_PREFIX=/home/boegelbot/easybuild/develop source init_env_easybuild_develop.sh; fi; EB_PR=23830 EB_ARGS= EB_CONTAINER= EB_REPO=easybuild-easyconfigs EB_BRANCH=develop /opt/software/slurm/bin/sbatch --job-name test_PR_23830 --ntasks=8 --partition=jsczen3g --gres=gpu:1 ~/boegelbot/eb_from_pr_upload_jsc-zen3.sh' executed!

exit code: 0
output:

Submitted batch job 7920

Test results coming soon (I hope)...

Details

- notification for comment with ID 3275022689 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

boegelbot · 2025-09-10T13:49:37Z

Test report by @boegelbot
FAILED
Build succeeded for 1 out of 2 (2 easyconfigs in total)
jsczen3g1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.6, x86_64, AMD EPYC-Milan Processor (zen3), 1 x NVIDIA NVIDIA A100 80GB PCIe, 575.57.08, Python 3.9.21
See https://gist.github.com/boegelbot/5c3002200a2328ef0f0b611f6cbd60fd for a full test report.

Co-authored-by: Simon Branford <4967+branfosj@users.noreply.github.com>

branfosj · 2025-09-11T08:54:11Z

Test report by @branfosj
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
bear-pg0105u03a - Linux RHEL 8.10, x86_64, Intel(R) Xeon(R) Platinum 8360Y CPU @ 2.40GHz (icelake), Python 3.6.8
See https://gist.github.com/branfosj/8cd9c47c4504fdc60fcef782bda01190 for a full test report.

Thyre · 2025-09-11T08:56:33Z

I created easybuilders/easybuild-framework#4996 to keep track of how we could improve the rpath sanity check.

branfosj · 2025-09-11T08:56:43Z

Test report by @branfosj
FAILED
Build succeeded for 0 out of 1 (1 easyconfigs in total)
bear-pg0208u23a - Linux RHEL 8.10, x86_64, Intel(R) Xeon(R) Platinum 8360Y CPU @ 2.40GHz (icelake), 1 x NVIDIA NVIDIA A100-SXM4-40GB, 560.35.05, Python 3.6.8
See https://gist.github.com/branfosj/0ff607d4f7e35428e8d00ee6aefbf118 for a full test report.

branfosj · 2025-09-11T09:05:34Z

Recording the test failures when performance counters are not available here, so that others can see them without having to search the logs.

jsc-zen3 has 6 failures:

hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC sampling produces profiles (gpu=nvidia)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC + tracing produces profiles (gpu=nvidia)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC + boosted tracing produces profiles (gpu=nvidia)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC sampling produces profiles (gpu=cuda)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC + tracing produces profiles (gpu=cuda)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC + boosted tracing produces profiles (gpu=cuda)

I have 12 - the above 6 and also:

hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ tracing produces profiles (gpu=cuda)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ boosted tracing produces profiles (gpu=cuda)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ boosted tracing produces profiles (gpu=nvidia)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda produces profiles (gpu=nvidia)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda produces profiles (gpu=cuda)
hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ tracing produces profiles (gpu=nvidia)

The ordering of the tests changes, but these pass on jsc-zen3 (tests 57-63). Not sure what difference causes the extra 6 to fail. I know my system has the relevant performance counters disabled.

Thyre · 2025-09-11T09:07:29Z

Recording the test failures when performance counters are not available here, so that others can see them without having to search the logs.

jsc-zen3 has 6 failures:

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC sampling produces profiles (gpu=nvidia)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC + tracing produces profiles (gpu=nvidia)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC + boosted tracing produces profiles (gpu=nvidia)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC sampling produces profiles (gpu=cuda)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC + tracing produces profiles (gpu=cuda)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ PC + boosted tracing produces profiles (gpu=cuda)`

I have 12 - the above 6 and also:

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ tracing produces profiles (gpu=cuda)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ boosted tracing produces profiles (gpu=cuda)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ boosted tracing produces profiles (gpu=nvidia)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda produces profiles (gpu=nvidia)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda produces profiles (gpu=cuda)`

* `hpctoolkit:hpcrun+cuda / Measurement of tstexe-vecadd-cuda w/ tracing produces profiles (gpu=nvidia)`

The ordering of the tests changes, but these pass on jsc-zen3 (tests 57-63). Not sure what difference causes the extra 6 to fail. I know my system has the relevant performance counters disabled.

The other 6 are very likely caused by the CUDA driver. We also see that with 570 on JUPITER, but don't see it on our other GH200 nodes with 580.

branfosj · 2025-09-11T09:12:36Z

The other 6 are very likely caused by the CUDA driver. We also see that with 570 on JUPITER, but don't see it on our other GH200 nodes with 580.

We have 560.35.05 so that fits the pattern.

Micket · 2025-09-15T09:22:22Z

Test report by @Micket
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
vera-r07-03 - Linux Rocky Linux 9.6, x86_64, Intel(R) Xeon(R) Gold 6338 CPU @ 2.00GHz, 4 x NVIDIA NVIDIA A40, 580.65.06, Python 3.9.21
See https://gist.github.com/Micket/a6b70f0150495b9f148cf4d8c327afb6 for a full test report.

Micket · 2025-09-15T09:33:06Z

Test report by @Micket
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
vera-r07-03 - Linux Rocky Linux 9.6, x86_64, Intel(R) Xeon(R) Gold 6338 CPU @ 2.00GHz, 4 x NVIDIA NVIDIA A40, 580.65.06, Python 3.9.21
See https://gist.github.com/Micket/66c37f3c04d631ad3ac1aef3438be6a2 for a full test report.

Micket · 2025-09-15T09:35:42Z

I'd like to get this merged.
While we could spend time, writing a custom easyblock to exclude tests that fail.. maybe they should just fail? It would tell the person that in order to fully use this software, they should update their drivers, or need to consider enabling performance counters. If not, they can always opt to accept the limitations and ignore the failing tests?

Anyone strongly disagrees?

Thyre · 2025-09-15T09:44:09Z

I'd like to get this merged. While we could spend time, writing a custom easyblock to exclude tests that fail.. maybe they should just fail? It would tell the person that in order to fully use this software, they should update their drivers, or need to consider enabling performance counters. If not, they can always opt to accept the limitations and ignore the failing tests?

Anyone strongly disagrees?

I think with the build message that now appears, users are informed enough to either work around the problem or install it with --ignore-test-failure knowing that some features might not work fully. We can certainly clean this up a bit more, but the EasyConfig is still readable enough that we can merge this as an EasyConfig only for now I'd say.

Micket

lgtm

Micket · 2025-09-16T14:32:39Z

Going in, thanks @Thyre!

adding easyconfigs: HPCToolkit-2025.0.0-gompi-2025b.eb, HPCToolkit-20…

a4fc616

…25.0.0-gompi-2025b-CUDA-12.9.1.eb

Thyre added the new label Sep 9, 2025

Thyre marked this pull request as draft September 9, 2025 12:43

Thyre added the 2025b issues & PRs related to 2025b common toolchains label Sep 9, 2025

github-actions Bot added update and removed new labels Sep 9, 2025

Thyre commented Sep 9, 2025

View reviewed changes

Comment thread easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.0-gompi-2025b.eb Outdated

Thyre and others added 6 commits September 9, 2025 14:51

Disable CUDA for non-CUDA HPCToolkit-2025.0.0-gompi-2025b

08314fd

Disable experimental Python bindings of HPCToolkit

bde723c

Fail tests due to incorrect values Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Enable tests of HPCToolkit

0af558d

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Disable OpenCL in HPCToolkit

1b21672

May pick up system OpenCL headers. Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Expand sanity check for HPCToolkit + CUDA

3cd836a

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Fix style for HPCToolkit EasyConfigs

6ff9855

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Cleanup HPCToolkit dependencies

13b4671

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Thyre added 2 commits September 9, 2025 16:34

Add note to tests that they can fail in certain scenarios

0efde03

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Merge remote-tracking branch 'upstream/develop' into 20250909144251_n…

ed352e2

…ew_pr_HPCToolkit202500

Thyre marked this pull request as ready for review September 9, 2025 14:35

Fix HPCToolkit sanity check style

b2907f7

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Thyre added 2 commits September 10, 2025 14:41

Update to 2025.0.1

07756c9

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Update checksums for new HPCToolkit version

e3cb9c7

Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

Thyre changed the title ~~{perf}[gompi/2025b] HPCToolkit v2025.0.0 w/ CUDA 12.9.1~~ {perf}[gompi/2025b] HPCToolkit v2025.0.1 w/ CUDA 12.9.1 Sep 10, 2025

Disable rpath sanity check due to preload libraries having rpath

38970f0

stripped See https://gitlab.com/hpctoolkit/hpctoolkit/-/merge_requests/1310 for more information on why this is done. Signed-off-by: Jan André Reuter <j.reuter@fz-juelich.de>

branfosj reviewed Sep 11, 2025

View reviewed changes

Comment thread easybuild/easyconfigs/h/HPCToolkit/HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb Outdated

Add build info message to HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1

a6825e2

Co-authored-by: Simon Branford <4967+branfosj@users.noreply.github.com>

Thyre mentioned this pull request Sep 11, 2025

Allow filtering of rpath sanity check in EasyConfigs / EasyBlocks easybuilders/easybuild-framework#4996

Open

Micket added this to the release after 5.1.2 (5.2.0?) milestone Sep 16, 2025

Micket approved these changes Sep 16, 2025

View reviewed changes

Micket merged commit 9c082a7 into easybuilders:develop Sep 16, 2025
8 checks passed

boegel modified the milestones: release after 5.1.2 (5.2.0?), next release (5.1.2) Sep 23, 2025

Thyre mentioned this pull request Oct 15, 2025

{tools}[GCCcore/14.3.0] HPCViewer v2025.2.0 #24249

Merged

Conversation

Thyre commented Sep 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented Sep 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Updated software HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb

Updated software HPCToolkit-2025.0.1-gompi-2025b.eb

Uh oh!

Uh oh!

Thyre commented Sep 9, 2025

Uh oh!

Thyre commented Sep 9, 2025

Uh oh!

Thyre commented Sep 9, 2025

Uh oh!

Thyre commented Sep 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Thyre commented Sep 9, 2025

Uh oh!

boegelbot commented Sep 9, 2025

Uh oh!

Thyre commented Sep 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Thyre commented Sep 9, 2025

Uh oh!

Thyre commented Sep 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

boegelbot commented Sep 9, 2025

Uh oh!

Thyre commented Sep 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Thyre commented Sep 10, 2025

Uh oh!

Thyre commented Sep 10, 2025

Uh oh!

Thyre commented Sep 10, 2025

Uh oh!

Thyre commented Sep 10, 2025

Uh oh!

boegelbot commented Sep 10, 2025

Uh oh!

boegelbot commented Sep 10, 2025

Uh oh!

Uh oh!

branfosj commented Sep 11, 2025

Uh oh!

Thyre commented Sep 11, 2025

Uh oh!

branfosj commented Sep 11, 2025

Uh oh!

branfosj commented Sep 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Thyre commented Sep 11, 2025

Uh oh!

branfosj commented Sep 11, 2025

Uh oh!

Micket commented Sep 15, 2025

Uh oh!

Micket commented Sep 15, 2025

Uh oh!

Micket commented Sep 15, 2025

Uh oh!

Thyre commented Sep 15, 2025

Uh oh!

Micket left a comment

Choose a reason for hiding this comment

Uh oh!

Micket commented Sep 16, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Thyre commented Sep 9, 2025 •

edited

Loading

github-actions Bot commented Sep 9, 2025 •

edited

Loading

Updated software `HPCToolkit-2025.0.1-gompi-2025b-CUDA-12.9.1.eb`

Updated software `HPCToolkit-2025.0.1-gompi-2025b.eb`

Thyre commented Sep 9, 2025 •

edited

Loading

Thyre commented Sep 9, 2025 •

edited

Loading

Thyre commented Sep 9, 2025 •

edited

Loading

Thyre commented Sep 9, 2025 •

edited

Loading

branfosj commented Sep 11, 2025 •

edited

Loading