Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

Remove 'make -f cudacpp -pn' to determine CUDACPP_BUILDDIR in madevent makefile #829

Closed
valassi opened this issue Mar 26, 2024 · 1 comment · Fixed by #798
Closed

Remove 'make -f cudacpp -pn' to determine CUDACPP_BUILDDIR in madevent makefile #829

valassi opened this issue Mar 26, 2024 · 1 comment · Fixed by #798
Assignees

Comments

@valassi
Copy link
Member

valassi commented Mar 26, 2024

While integrating Jorgen's makefile changes for separate targets in PR #798, I am removing the 'make -f cudacpp -pn' to determine CUDACPP_BUILDDIR in madevent makefile. This is one of the most horrible hacks I had included in the makefiles.

For this one, I am just adding a separate makefile included by both cudacpp.mk an dthe fortran makefile. This was one of the options we discussed in December. This is essentially done. Filing this for reference, cross linked with the relevant changes.

@valassi valassi self-assigned this Mar 26, 2024
@valassi valassi linked a pull request Mar 26, 2024 that will close this issue
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
Note: I noticed that building madevent for HIP on a CUDA node gives a weird error in the fortran makefile.
I will start cleaning up the way the CUDACPP_BUILDDIR is determined in the makefile madgraph5#829.

make -j madevent_hip_link
  CUDACPP_BUILDDIR='.'
  make BACKEND=hip ./madevent_hip
  make[1]: Entering directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx'
  makefile:62: *** CUDACPP_BUILDDIR='' should not be empty!.  Stop.
  make[1]: Leaving directory '/data/avalassi/GPU2023/madgraph4gpuX/epochX/cudacpp/gg_tt.mad/SubProcesses/P1_gg_ttx'
  make: *** [makefile:170: madevent_hip_link] Error 2
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
… BACKEND (start cleaning up CUDACPP_BUILDDIR in fortran makefile madgraph5#829)
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
…-defined FPTYPE, HELINL, HRDCOD and the choice of a default one madgraph5#829
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
… a new CUDACPP_BUILDDIR (later used to define BUILDDIR itself) madgraph5#829
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
…udacpp -pn' to determine CUDACPP_BUILDDIR in madevent makefile madgraph5#829

./CODEGEN/generateAndCompare.sh gg_tt --mad --nopatch
git diff --no-ext-diff -R gg_tt.mad/Source/makefile gg_tt.mad/Source/dsample.f gg_tt.mad/Source/genps.inc gg_tt.mad/SubProcesses/makefile > CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/MG5aMC_patches/PROD/patch.common
git diff --no-ext-diff -R gg_tt.mad/bin/internal/banner.py gg_tt.mad/bin/internal/gen_ximprove.py gg_tt.mad/bin/internal/madevent_interface.py >> CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/MG5aMC_patches/PROD/patch.common
git diff --no-ext-diff -R gg_tt.mad/SubProcesses/P1_gg_ttx/auto_dsig1.f gg_tt.mad/SubProcesses/P1_gg_ttx/driver.f gg_tt.mad/SubProcesses/P1_gg_ttx/matrix1.f > CODEGEN/PLUGIN/CUDACPP_SA_OUTPUT/MG5aMC_patches/PROD/patch.P1
git checkout gg_tt.mad
valassi added a commit to valassi/madgraph4gpu that referenced this issue Mar 26, 2024
…r HIP_HOME externally

This is now no longer necessary, as in any case CUDA and C++ builds are all done separately
(and it is possible to disable CURAND and even NVTX if needed)

Note: the check whether hipcc and nvcc are found for BACKEND=hop adn cuda is moved to cudacpp_builddir.mk madgraph5#829
@valassi
Copy link
Member Author

valassi commented Mar 26, 2024

This is completed in PR #798. There is a new cudacpp_builddir.mk file.

# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

1 participant