Skip to content
New issue

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

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

Already on GitHub? Sign in to your account

if we are building with HIP, don't force inlining #1490

Merged
merged 2 commits into from
Feb 22, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions .github/workflows/hip.yml
Original file line number Diff line number Diff line change
Expand Up @@ -33,10 +33,10 @@ jobs:
run: |
cd unit_test/test_react
# -amdgpu-function-calls=true tells the compiler not to force-inline everything
make USE_HIP=TRUE USE_MPI=FALSE USE_OMP=FALSE USE_CUDA=FALSE NETWORK_DIR=iso7 EXTRACXXFLAGS='-mllvm -amdgpu-function-calls=true' -j 4
make USE_HIP=TRUE USE_MPI=FALSE USE_OMP=FALSE USE_CUDA=FALSE NETWORK_DIR=iso7 -j 4

- name: compile test_react with HIP (ignition_reaclib/URCA-simple)
run: |
cd unit_test/test_react
make realclean
make USE_HIP=TRUE USE_MPI=FALSE USE_OMP=FALSE USE_CUDA=FALSE NETWORK_DIR=ignition_reaclib/URCA-simple EXTRACXXFLAGS='-mllvm -amdgpu-function-calls=true' -j 4
make USE_HIP=TRUE USE_MPI=FALSE USE_OMP=FALSE USE_CUDA=FALSE NETWORK_DIR=ignition_reaclib/URCA-simple -j 4
5 changes: 5 additions & 0 deletions Make.Microphysics_extern
Original file line number Diff line number Diff line change
@@ -1,6 +1,11 @@
# This is the main include makefile for applications that want to use Microphysics
# You should set NETWORK_OUTPUT_PATH before including this file

# For HIP, inlining produces large kernels that seem to cause memory issues
ifeq ($(USE_HIP),TRUE)
EXTRACXXFLAGS += -mllvm -amdgpu-function-calls=true
endif

# for common code paths between simplified-SDC and true-SDC, we'll
# use the SDC define

Expand Down