Fix cudacpp makefile for installations (e.g. in CMS) where nvcc exists but curand and nvtx headers are missing #966
Merged
valassi merged 6 commits intomadgraph5:masterfrom Aug 21, 2024
Merged
Conversation
…f 'USE_NVTX=' is defined
…e not found (madgraph5#965), also clean up NVTX and debug printouts
…AND if NVTX/CURAND headers are not found (madgraph5#965) Also remove the option to disable NVTX if 'USE_NVTX='. Keep the option to disable CURAND if 'HASCURAND=hasNoCurand'.
This was referenced Aug 17, 2024
Member
Author
|
Hi @oliviermattelaer can you review this one? It fixes the issue #965, @choij1589 confirms it is good. (Note, some CI tests are failing because of Mac configuration issues #971... orthogonal to this issue) Thanks |
oliviermattelaer
approved these changes
Aug 20, 2024
Member
oliviermattelaer
left a comment
oliviermattelaer
left a comment
There was a problem hiding this comment.
This can move forward (even if it is indeed a strong argument to move to a more automatic/flexible way to compile the code, like you said
Member
Author
Member
Author
Thanks @oliviermattelaer ! Ok so I am merging this one. |
valassi
added a commit
to valassi/madgraph4gpu
that referenced
this pull request
Aug 21, 2024
…MS nvcc without nvtx PR madgraph5#966) into hel
valassi
added a commit
to valassi/madgraph4gpu
that referenced
this pull request
Aug 21, 2024
… mac madgraph5#974, nvcc madgraph5#966) into june24 Fix conflicts: epochX/cudacpp/CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/MG5aMC_patches/PROD/patch.P1 epochX/cudacpp/CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/madgraph/iolibs/template_files/gpu/counters.cc epochX/cudacpp/CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/madgraph/iolibs/template_files/gpu/fbridge.cc epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/auto_dsig1.f epochX/cudacpp/gg_tt.mad/SubProcesses/counters.cc epochX/cudacpp/gg_tt.mad/SubProcesses/fbridge.cc NB: here I essentially fixed gg_tt.mad, not CODEGEN, which will need to be adjusted a posteriori with a backport In particular: - Note1: patch.P1 is now taken from june24, but will need to be recomputed git checkout HEAD CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/MG5aMC_patches/PROD/patch.P1 - Note2: I need to manually port some upstream/master changes in auto_dsig1.f to smatrix_multi.f, which did not yet exist
valassi
added a commit
to valassi/madgraph4gpu
that referenced
this pull request
Aug 21, 2024
… mac madgraph5#974, nvcc madgraph5#966) into pptt Fix conflicts: epochX/cudacpp/CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/MG5aMC_patches/PROD/patch.P1 (take HEAD version, must recompute) epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx/auto_dsig1.f (fix manually)
valassi
added a commit
to valassi/madgraph4gpu
that referenced
this pull request
Aug 21, 2024
… mac madgraph5#974, nvcc madgraph5#966) into prof
valassi
added a commit
to valassi/madgraph4gpu
that referenced
this pull request
Aug 22, 2024
… mac madgraph5#974, nvcc madgraph5#966) into grid
valassi
added a commit
to valassi/madgraph4gpu
that referenced
this pull request
Aug 22, 2024
…dgraph5#960, mac madgraph5#974, nvcc madgraph5#966) into cmsdy Fix conflict in tlau/fromgridpacks/parseGridpackLogs.sh (use the currenmt cmsdy version: git checkout b125b65 tlau/fromgridpacks/parseGridpackLogs.sh)
valassi
added a commit
to valassi/madgraph4gpu
that referenced
this pull request
Aug 22, 2024
…rge with hel madgraph5#960, mac madgraph5#974, nvcc madgraph5#966) into cmsdyps
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Hi @oliviermattelaer this is a very simple PR, which only modifies the cudacpp.mk for installations (e.g. in CMS) where nvcc exists but curand and nvtx headers are missing.
It fixes the issues reported by @choij1589 in #965 and at the meeting yesterday.
Jin could already work around the issue by setting
USE_NVTX= HASCURAND=hasNoCurandbut I think that t is much better if some sanity checks for the existence of the headers is done elsewhere. (Yes I know cmake would make that out of the box, but we are not there).Can you have a look please? This is rather standalone and independent from all other PRs I submitted.
Thanks
Andrea