mirrors/rocm_jax
1
0
Fork 0
mirror of https://github.com/ROCm/jax.git synced 2025-04-14 10:56:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

Update the calculation for num_processes and num_test_jobs that are used in CUDA test jobs

Browse Source 
We need to set them as `min(num_cpu_cores, num_gpus * max_tests_per_gpu, total ram in GB/6)` where max_tests_per_gpu = (GPU memory / 2GB)

PiperOrigin-RevId: 731730857
This commit is contained in:
Nitin Srinivasan 2025-02-27 08:10:58 -08:00 committed by jax authors
parent 4997e45743
commit 5ae0e58a4a
3 changed files with 43 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
ci/envs/default.env
View File

@ -53,13 +53,6 @@ export JAXCI_WRITE_TO_BAZEL_REMOTE_CACHE=${JAXCI_WRITE_TO_BAZEL_REMOTE_CACHE:-0}
# #############################################################################
# Test script specific environment variables.
# #############################################################################
# The maximum number of tests to run per GPU when running single accelerator
# tests with parallel execution with Bazel. The GPU limit is set because we
# need to allow about 2GB of GPU RAM per test. Default is set to 12 because we
# use L4 machines which have 24GB of RAM but can be overriden if we use a
# different GPU type.
export JAXCI_MAX_TESTS_PER_GPU=${JAXCI_MAX_TESTS_PER_GPU:-12}
# Sets the value of `JAX_ENABLE_X64` in the test scripts. CI builds override
# this value in the Github action workflow files.
export JAXCI_ENABLE_X64=${JAXCI_ENABLE_X64:-0}

24
ci/run_bazel_test_cuda_non_rbe.sh
View File

@ -37,14 +37,30 @@ nvidia-smi
echo "Running single accelerator tests (without RBE)..."
# Set up test environment variables.
# Set the number of test jobs to min(num_cpu_cores, gpu_count * max_tests_per_gpu, total_ram_gb / 6)
# We calculate max_tests_per_gpu as memory_per_gpu_gb / 2gb
# Calculate gpu_count * max_tests_per_gpu
export gpu_count=$(nvidia-smi --query-gpu=name --format=csv,noheader | wc -l)
export num_test_jobs=$((gpu_count * JAXCI_MAX_TESTS_PER_GPU))
export memory_per_gpu_gb=$(nvidia-smi --query-gpu=memory.total --format=csv,noheader,nounits --id=0)
export memory_per_gpu_gb=$((memory_per_gpu_gb / 1024))
# Allow 2 GB of GPU RAM per test
export max_tests_per_gpu=$((memory_per_gpu_gb / 2))
export num_test_jobs=$((gpu_count * max_tests_per_gpu))
# Calculate num_cpu_cores
export num_cpu_cores=$(nproc)
# tests_jobs = max(gpu_count * max_tests_per_gpu, num_cpu_cores)
if [[ $num_test_jobs -gt $num_cpu_cores ]]; then
# Calculate total_ram_gb / 6
export total_ram_gb=$(awk '/MemTotal/ {printf "%.0f", $2/1048576}' /proc/meminfo)
export host_memory_limit=$((total_ram_gb / 6))
if [[ $num_cpu_cores -lt $num_test_jobs ]]; then
num_test_jobs=$num_cpu_cores
fi
if [[ $host_memory_limit -lt $num_test_jobs ]]; then
num_test_jobs=$host_memory_limit
fi
# End of test environment variables setup.
# Don't abort the script if one command fails to ensure we run both test
@ -64,7 +80,7 @@ bazel test --config=ci_linux_x86_64_cuda \
--run_under "$(pwd)/build/parallel_accelerator_execute.sh" \
--test_output=errors \
--test_env=JAX_ACCELERATOR_COUNT=$gpu_count \
--test_env=JAX_TESTS_PER_ACCELERATOR=$JAXCI_MAX_TESTS_PER_GPU \
--test_env=JAX_TESTS_PER_ACCELERATOR=$max_tests_per_gpu \
--local_test_jobs=$num_test_jobs \
--test_env=JAX_EXCLUDE_TEST_TARGETS=PmapTest.testSizeOverflow \
--test_tag_filters=-multiaccelerator \

25
ci/run_pytest_cuda.sh
View File

@ -45,9 +45,30 @@ export NCCL_DEBUG=WARN
export TF_CPP_MIN_LOG_LEVEL=0
export JAX_ENABLE_64="$JAXCI_ENABLE_X64"
# Set the number of processes to run to be 4x the number of GPUs.
# Set the number of processes to min(num_cpu_cores, gpu_count * $max_tests_per_gpu, total_ram_gb / 6)
# We calculate max_tests_per_gpu as memory_per_gpu_gb / 2gb
# Calculate gpu_count * max_tests_per_gpu
export gpu_count=$(nvidia-smi --query-gpu=name --format=csv,noheader | wc -l)
export num_processes=`expr 4 \* $gpu_count`
export memory_per_gpu_gb=$(nvidia-smi --query-gpu=memory.total --format=csv,noheader,nounits --id=0)
export memory_per_gpu_gb=$((memory_per_gpu_gb / 1024))
# Allow 2 GB of GPU RAM per test
export max_tests_per_gpu=$((memory_per_gpu_gb / 2))
export num_processes=$((gpu_count * max_tests_per_gpu))
# Calculate num_cpu_cores
export num_cpu_cores=$(nproc)
# Calculate total_ram_gb / 6
export total_ram_gb=$(awk '/MemTotal/ {printf "%.0f", $2/1048576}' /proc/meminfo)
export host_memory_limit=$((total_ram_gb / 6))
if [[ $num_cpu_cores -lt $num_processes ]]; then
num_processes=$num_cpu_cores
fi
if [[ $host_memory_limit -lt $num_processes ]]; then
num_processes=$host_memory_limit
fi
export XLA_PYTHON_CLIENT_ALLOCATOR=platform
export XLA_FLAGS=--xla_gpu_force_compilation_parallelism=1