{vis}[GCCcore/14.2.0] LLVM v20.1.7, Mesa v25.1.3, lit v18.1.8, ...#23144
{vis}[GCCcore/14.2.0] LLVM v20.1.7, Mesa v25.1.3, lit v18.1.8, ...#23144Crivella merged 9 commits intoeasybuilders:developfrom
Conversation
…e-14.2.0.eb, lit-18.1.8-GCCcore-14.2.0.eb, Xvfb-21.1.18-GCCcore-14.2.0.eb
Updated software
|
|
Test report by @Crivella |
|
Test report by @Crivella cmd.shError from re-running the failed command |
|
Updated easyblock fixing the long long long deprecated "swrast" option. |
|
Test report by @Crivella |
|
it needs the updated easyblock for mesa to remove the swrast flag, though i haven't tested any of this myself as my builds have been slow to catch up |
|
I used I think we need https://github.com/easybuilders/easybuild-easyblocks/pull/3790/files#r2158321213 added, for the wrong direction in the comparison If i do it manually i then hit Guess we are also gonna need a package for https://github.com/KhronosGroup/glslang ? |
|
I'm stuck on LLVM itself failing to build, 105 failed tests on my machine, so I'm struggling to test mesa itself. |
Co-authored-by: Davide Grassano <34096612+Crivella@users.noreply.github.com>
|
Test report by @Crivella Notes
|
|
I'm not getting past LLVM still. Builds take an eternity. Testing is also way to excessive. It takes a whole working day to just get to the failures. I tried using the latest llvm.py easyblock in case there was something there. This is on a standard intel icelake node with rocky linux 9.6. A whole bunch of tests are all timing out for me. |
|
The failing ones seems fairly standard and nothing to worry about, One thing you could try is bumping up the max timeout per single test. EDIT: alternatively we could add a flag to the easyblock to not consider timed out tests as failures to work around this |
|
Test report by @Thyre |
$ ls -lah /data/EasyBuild-develop/software/Mesa/25.1.3-GCCcore-14.2.0/include
total 48K
drwxr-xr-x 4 jreuter jreuter 4,0K 7. Jul 09:47 .
drwxr-xr-x 5 jreuter jreuter 4,0K 7. Jul 09:47 ..
drwxr-xr-x 2 jreuter jreuter 4,0K 7. Jul 09:47 EGL
-rw-r--r-- 1 jreuter jreuter 12K 7. Jun 15:47 gbm_backend_abi.h
-rw-r--r-- 1 jreuter jreuter 19K 7. Jun 15:47 gbm.h
drwxr-xr-x 3 jreuter jreuter 4,0K 7. Jul 09:47 GL
$ ls -lah /data/EasyBuild-develop/software/Mesa/25.1.3-GCCcore-14.2.0/lib
total 153M
drwxr-xr-x 5 jreuter jreuter 4,0K 7. Jul 09:47 .
drwxr-xr-x 5 jreuter jreuter 4,0K 7. Jul 09:47 ..
drwxr-xr-x 2 jreuter jreuter 4,0K 7. Jul 09:47 dri
drwxr-xr-x 2 jreuter jreuter 4,0K 7. Jul 09:47 gbm
lrwxrwxrwx 1 jreuter jreuter 16 7. Jul 09:47 libEGL_mesa.so -> libEGL_mesa.so.0
lrwxrwxrwx 1 jreuter jreuter 20 7. Jul 09:47 libEGL_mesa.so.0 -> libEGL_mesa.so.0.0.0
-rwxr-xr-x 1 jreuter jreuter 4,4M 7. Jul 09:47 libEGL_mesa.so.0.0.0
-rwxr-xr-x 1 jreuter jreuter 86M 7. Jul 09:47 libgallium-25.1.3.so
lrwxrwxrwx 1 jreuter jreuter 11 7. Jul 09:47 libgbm.so -> libgbm.so.1
lrwxrwxrwx 1 jreuter jreuter 15 7. Jul 09:47 libgbm.so.1 -> libgbm.so.1.0.0
-rwxr-xr-x 1 jreuter jreuter 363K 7. Jul 09:47 libgbm.so.1.0.0
lrwxrwxrwx 1 jreuter jreuter 16 7. Jul 09:47 libGLX_indirect.so.0 -> libGLX_mesa.so.0
lrwxrwxrwx 1 jreuter jreuter 16 7. Jul 09:47 libGLX_mesa.so -> libGLX_mesa.so.0
lrwxrwxrwx 1 jreuter jreuter 20 7. Jul 09:47 libGLX_mesa.so.0 -> libGLX_mesa.so.0.0.0
-rwxr-xr-x 1 jreuter jreuter 2,4M 7. Jul 09:47 libGLX_mesa.so.0.0.0
-rwxr-xr-x 1 jreuter jreuter 479K 7. Jul 09:47 libVkLayer_MESA_device_select.so
-rwxr-xr-x 1 jreuter jreuter 60M 7. Jul 09:47 libvulkan_lvp.so
drwxr-xr-x 2 jreuter jreuter 4,0K 7. Jul 09:47 pkgconfigAh okay, that failure is known (see #23144 (comment)). |
|
Test report by @Thyre |
|
LLVM 20.1.8 was released: https://github.com/llvm/llvm-project/releases/tag/llvmorg-20.1.8 |
Would probably keep |
Fine with me 👍 |
|
@boegelbot please test @ jsc-zen3 |
|
@Crivella: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de PR test command '
Test results coming soon (I hope)... Details- notification for comment with ID 3079131588 processed Message to humans: this is just bookkeeping information for me, |
|
@Crivella I think that's a good addition to the EasyBlock. I'm wondering: was your intention to then also set it in this EasyConfig? I personally wouldn't: it seems to be pretty specific to @Micket 's setup. Sites that encounter this could add it through a hook. And if we don't plan to add it to this EasyConfig, there's no point in delaying this PR for this (a fair amount of other PRs depend on it). I'll push the suggested changes from @Thyre myself (hope you're ok with that @Micket ) so that we can re-test this. Then, I think, we can finally merge this and proceed to the dependent PRs... |
|
My idea for easybuilders/easybuild-easyblocks#3903 would be to leave it always |
…w. Added zstd as dependency
Ok, perfect, in that case it doesn't have an impact on the EasyConfig for LLVM in this PR. I've pushed the small changes suggested by @Thyre and will trigger retesting. |
|
@boegelbot please test @ jsc-zen3 |
|
@casparvl: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de PR test command '
Test results coming soon (I hope)... Details- notification for comment with ID 3233126994 processed Message to humans: this is just bookkeeping information for me, |
|
Test report by @boegelbot |
|
Test report by @casparvl |
|
MB, ran this without pulling in the changes for the easyblock in my dev environment Test report by @Crivella |
|
Test report by @Crivella NOTEBuild with |
|
@boegelbot please test @ jsc-zen3 |
|
@Crivella: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de PR test command '
Test results coming soon (I hope)... Details- notification for comment with ID 3249040932 processed Message to humans: this is just bookkeeping information for me, |
|
@Crivella Where is this stuck currently? |
@boegel We need to decide if we want to go with the monolithic or split build for After one of them is merged we can move forward with this |
Seems like the monolitic one makes more sense, but we should rename to |
…asyconfigs into 20250619010219_new_pr_LLVM2017
|
@boegelbot please test @ jsc-zen3 |
|
@Crivella: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de PR test command '
Test results coming soon (I hope)... Details- notification for comment with ID 3275555482 processed Message to humans: this is just bookkeeping information for me, |
|
Test report by @boegelbot |
|
Test report by @Crivella NOTE: build launched after f7400b4 |
|
Going in, thanks @Micket! |
(created using
eb --new-pr)Completely untested, just bumping versions so far. To tired to continue tonight. Is this the right versions to use for LLVM here?
Requires
Edit by @casparvl : as far as I can determine, this now also depends on
swrastas gallium driver for Mesa versions >= 25.x easybuild-easyblocks#3790amdgcn_capabilitiesbuild option easybuild-easyblocks#3824