microsoft/mscclpp
1
0
Fork 0
mirror of https://github.com/microsoft/mscclpp.git synced 2026-05-11 17:00:22 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
85bcc9a8a6ccb11ba2d668000c54bb67fde4d5c1
mscclpp/test/deploy/run_tests.sh
Copilot 93f6eeaa6b Remove GTest dependency, add code coverage, and refactor unit tests and CI pipelines (#744) 
- Removes the GTest dependency, replacing it with a minimal custom
framework (`test/framework.*`) that covers only what the tests actually
use — a unified `TEST()` macro with SFINAE-based fixture auto-detection,
`EXPECT_*`/`ASSERT_*` assertions, environments, and setup/teardown.
- `--exclude-perf-tests` flag and substring-based negative filtering
- `MSCCLPP_ENABLE_COVERAGE` CMake option with gcov/lcov; CI uploads to
Codecov
- Merges standalone `test/perf/` into main test targets
- Refactors Azure pipelines to reduce redundancies & make more readable

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: Changho Hwang <changhohwang@microsoft.com>
2026-03-24 23:34:38 -04:00

110 lines
5.7 KiB
Bash
Raw Blame History

set -e
HOSTFILE=/root/mscclpp/test/deploy/hostfile_mpi
function run_mscclpp_test()
{
echo "=================Run allgather_test_perf on 2 nodes========================="
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/allgather_test_perf -b 1K -e 1G -f 2 -k 0 -o /root/mscclpp/output.jsonl
# For kernel 2, the message size must can be divided by 3
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/allgather_test_perf -b 3K -e 3G -f 2 -k 2 -o /root/mscclpp/output.jsonl
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/allgather_test_perf -b 1K -e 1G -f 2 -k 3 -o /root/mscclpp/output.jsonl
echo "==================Run allreduce_test_perf on 2 nodes========================="
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/allreduce_test_perf -b 1K -e 1G -f 2 -k 0 -o /root/mscclpp/output.jsonl
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/allreduce_test_perf -b 1K -e 1G -f 2 -k 1 -o /root/mscclpp/output.jsonl
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/allreduce_test_perf -b 1K -e 1M -f 2 -k 2 -o /root/mscclpp/output.jsonl
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/allreduce_test_perf -b 3K -e 3G -f 2 -k 3 -o /root/mscclpp/output.jsonl
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/allreduce_test_perf -b 3K -e 3G -f 2 -k 4 -o /root/mscclpp/output.jsonl
echo "==================Run alltoall_test_perf on 2 nodes========================="
mpirun --allow-run-as-root -np 16 --bind-to numa -hostfile ${HOSTFILE} \
-x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mscclpp-test/alltoall_test_perf -b 1K -e 1G -f 2 -k 0 -o /root/mscclpp/output.jsonl
echo "========================Run performance check==============================="
python3 /root/mscclpp/test/mscclpp-test/check_perf_result.py --perf-file /root/mscclpp/output.jsonl \
--baseline-file /root/mscclpp/test/deploy/perf_ndmv4.jsonl
}
function run_mp_ut()
{
echo "============Run multi-process unit tests on 2 nodes (np=2, npernode=1)========================="
mpirun -allow-run-as-root -tag-output -np 2 --bind-to numa \
-hostfile ${HOSTFILE} -x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 1 /root/mscclpp/build/test/mp_unit_tests -ip_port mscclit-000000:20003
echo "============Run multi-process unit tests on 2 nodes (np=16, npernode=8)========================="
mpirun -allow-run-as-root -tag-output -np 16 --bind-to numa \
-hostfile ${HOSTFILE} -x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-npernode 8 /root/mscclpp/build/test/mp_unit_tests -ip_port mscclit-000000:20003
}
function run_pytests()
{
echo "==================Run python tests================================"
mpirun -allow-run-as-root -tag-output -np 16 --bind-to numa \
-hostfile ${HOSTFILE} -x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-x MSCCLPP_HOME=/root/mscclpp -npernode 8 bash /root/mscclpp/test/deploy/pytest.sh
}
function run_py_benchmark()
{
echo "==================Run python benchmark================================"
mpirun -allow-run-as-root -np 16 --bind-to numa \
-hostfile ${HOSTFILE} -x MSCCLPP_DEBUG=WARN -x LD_LIBRARY_PATH=/root/mscclpp/build/lib:$LD_LIBRARY_PATH \
-mca pml ob1 -mca btl ^openib -mca btl_tcp_if_include eth0 -x NCCL_IB_PCI_RELAXED_ORDERING=1 -x NCCL_SOCKET_IFNAME=eth0 \
-x CUDA_DEVICE_ORDER=PCI_BUS_ID -x NCCL_NET_GDR_LEVEL=5 -x NCCL_TOPO_FILE=/opt/microsoft/ndv4-topo.xml \
-x NCCL_NET_PLUGIN=none -x NCCL_IB_DISABLE=0 -x NCCL_MIN_NCHANNELS=32 -x NCCL_DEBUG=WARN -x NCCL_P2P_DISABLE=0 -x NCCL_SHM_DISABLE=0 \
-x MSCCLPP_HOME=/root/mscclpp -np 16 -npernode 8 python3 /root/mscclpp/python/mscclpp_benchmark/allreduce_bench.py
}
if [ $# -lt 1 ]; then
echo "Usage: $0 <mscclpp-test/mp-ut/run_pytests/run_py_benchmark>"
exit 1
fi
test_name=$1
case $test_name in
mscclpp-test)
echo "==================Run mscclpp-test on 2 nodes========================="
run_mscclpp_test
;;
mp-ut)
echo "==================Run mp-ut on 2 nodes================================"
run_mp_ut
;;
pytests)
echo "==================Run python tests===================================="
run_pytests
;;
py-benchmark)
echo "==================Run python benchmark================================"
run_py_benchmark
;;
*)
echo "Unknown test name: $test_name"
exit 1
;;
esac
Reference in New Issue View Git Blame Copy Permalink