Skip to content

Toolchain: Workaround incorrect ELPA results with GCC 15.2 #7784

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.

Sign up for GitHub

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

Open
Growl1234 wants to merge 1 commit into
base: develop
Choose a base branch
from
+13 −2
Open

Toolchain: Workaround incorrect ELPA results with GCC 15.2 #7784

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
15 changes: 13 additions & 2 deletions toolchain/scripts/stage3/install_elpa.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,6 +109,17 @@ case "$with_elpa" in
config_flags="--enable-avx-kernels=${has_AVX} --enable-avx2-kernels=${has_AVX2} --enable-avx512-kernels=${has_AVX512}"
fi
fi

# GCC 15.2 miscompiles ELPA's double-complex AVX/AVX2 BLOCK2
# kernels when tree SLP vectorization is enabled.
elpa_c_workaround_flags=""
cc_version="$("${MPICC}" -dumpfullversion -dumpversion 2>/dev/null || true)"
cc_banner="$("${MPICC}" --version 2>/dev/null | head -n 1 || true)"
if [[ "${cc_banner}" == *gcc* || "${cc_banner}" == *GCC* ]] &&
[[ "${cc_version}" =~ ^15\.2(\.|$) ]]; then
elpa_c_workaround_flags="-fno-tree-slp-vectorize"
fi

for TARGET in "cpu" "nvidia"; do
# Accept both uppercase and lowercase GPU enable flags for compatibility
gpu_enabled="${ENABLE_CUDA}"
Expand Down Expand Up @@ -140,7 +151,7 @@ case "$with_elpa" in
CXX=${MPICXX} \
CPP="cpp -E" \
FCFLAGS="${FCFLAGS} ${MATH_CFLAGS} ${SCALAPACK_CFLAGS} ${AVX_flag} ${FMA_flag} ${SSE4_flag} ${AVX512_flags} -fno-lto" \
CFLAGS="${CFLAGS} ${MATH_CFLAGS} ${SCALAPACK_CFLAGS} ${AVX_flag} ${FMA_flag} ${SSE4_flag} ${AVX512_flags} -fno-lto" \
CFLAGS="${CFLAGS} ${MATH_CFLAGS} ${SCALAPACK_CFLAGS} ${AVX_flag} ${FMA_flag} ${SSE4_flag} ${AVX512_flags} -fno-lto ${elpa_c_workaround_flags}" \
CXXFLAGS="${CXXFLAGS} ${MATH_CFLAGS} ${SCALAPACK_CFLAGS} ${AVX_flag} ${FMA_flag} ${SSE4_flag} ${AVX512_flags} -fno-lto" \
LDFLAGS="${MATH_LDFLAGS} ${SCALAPACK_LDFLAGS} ${cray_ldflags} -lstdc++" \
LIBS="${SCALAPACK_LIBS} $(resolve_string "${MATH_LIBS}" "MPI") ${MPI_LIBS}" \
Expand Down Expand Up @@ -171,7 +182,7 @@ case "$with_elpa" in
CXX=${MPICXX} \
CPP="cpp -E" \
FCFLAGS="${FCFLAGS} ${MATH_CFLAGS} ${SCALAPACK_CFLAGS} ${AVX_flag} ${FMA_flag} ${SSE4_flag} ${AVX512_flags} -fno-lto" \
CFLAGS="${CFLAGS} ${MATH_CFLAGS} ${SCALAPACK_CFLAGS} ${AVX_flag} ${FMA_flag} ${SSE4_flag} ${AVX512_flags} -fno-lto" \
CFLAGS="${CFLAGS} ${MATH_CFLAGS} ${SCALAPACK_CFLAGS} ${AVX_flag} ${FMA_flag} ${SSE4_flag} ${AVX512_flags} -fno-lto ${elpa_c_workaround_flags}" \
CXXFLAGS="${CXXFLAGS} ${MATH_CFLAGS} ${SCALAPACK_CFLAGS} ${AVX_flag} ${FMA_flag} ${SSE4_flag} ${AVX512_flags} -fno-lto" \
LDFLAGS="-Wl,--allow-multiple-definition -Wl,--enable-new-dtags ${MATH_LDFLAGS} ${SCALAPACK_LDFLAGS} ${cray_ldflags} -lstdc++" \
LIBS="${SCALAPACK_LIBS} $(resolve_string "${MATH_LIBS}" "MPI")" \
Expand Down
Loading