LLM inference in C/C++
C++ ggml
ci : cap test-backend-ops parallel jobs at 2 and add a 3600s timeout (#28833)
Georgi Gerganov002a12ad2550
diff --git a/ci/run.sh b/ci/run.shindex 1ceb19fd50..e510b8c45e 100755--- a/ci/run.sh+++ b/ci/run.sh@@ -775,7 +775,11 @@ function gg_run_test_backend_ops {set -e- local args_extra="-j $(nproc)"+ local n_jobs=$(nproc)+ if [ "${n_jobs}" -gt 2 ]; then+ n_jobs=2+ fi+ local args_extra="-j ${n_jobs}"# TODO: fix multi-threaded for ROCm# https://github.com/ggml-org/llama.cpp/actions/runs/34576278519/job/103297889044?pr=28740#step:3:4865@@ -789,10 +793,11 @@ function gg_run_test_backend_ops {args_extra=""fi+ # TODO: reduce the test-backend-ops timeout to 1800sif [ ! -z ${GG_BUILD_HIGH_PERF} ]; then- (time ./bin/test-backend-ops ${args_extra} -b CPU) 2>&1 | tee -a $OUT/${ci}-test-backend-ops.log+ (time timeout 3600 ./bin/test-backend-ops ${args_extra} -b CPU) 2>&1 | tee -a $OUT/${ci}-test-backend-ops.logelse- (time ./bin/test-backend-ops ${args_extra} ) 2>&1 | tee -a $OUT/${ci}-test-backend-ops.log+ (time timeout 3600 ./bin/test-backend-ops ${args_extra} ) 2>&1 | tee -a $OUT/${ci}-test-backend-ops.logfiset +e