ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

Update oneccl-binding.patch (#12377)

Browse source 
* Add files via upload

* upload oneccl-binding.patch

* Update Dockerfile
This commit is contained in:
Shaojun Liu 2024-11-11 22:34:08 +08:00 committed by GitHub
parent e0918934c8
commit c92d76b997
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 26 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
docker/llm/serving/xpu/docker/Dockerfile
View file

@ -10,6 +10,7 @@ ENV PYTHONUNBUFFERED=1
# Disable pip's cache behavior # Disable pip's cache behavior
ARG PIP_NO_CACHE_DIR=false ARG PIP_NO_CACHE_DIR=false
ADD ./gradio_web_server.patch /tmp/gradio_web_server.patch ADD ./gradio_web_server.patch /tmp/gradio_web_server.patch
ADD ./oneccl-binding.patch /tmp/oneccl-binding.patch
RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor | tee /usr/share/keyrings/intel-oneapi-archive-keyring.gpg > /dev/null && \ RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor | tee /usr/share/keyrings/intel-oneapi-archive-keyring.gpg > /dev/null && \
echo "deb [signed-by=/usr/share/keyrings/intel-oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main " | tee /etc/apt/sources.list.d/oneAPI.list && \ echo "deb [signed-by=/usr/share/keyrings/intel-oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main " | tee /etc/apt/sources.list.d/oneAPI.list && \
@ -57,6 +58,7 @@ RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRO
bash oneccl_wks_installer_2024.0.0.5.1.sh && \ bash oneccl_wks_installer_2024.0.0.5.1.sh && \
git clone https://github.com/intel/torch-ccl -b v2.1.300+xpu && \ git clone https://github.com/intel/torch-ccl -b v2.1.300+xpu && \
cd torch-ccl && \ cd torch-ccl && \
patch -p1 < /tmp/oneccl-binding.patch && \
USE_SYSTEM_ONECCL=ON COMPUTE_BACKEND=dpcpp python setup.py install && \ USE_SYSTEM_ONECCL=ON COMPUTE_BACKEND=dpcpp python setup.py install && \
apt-get update && \ apt-get update && \
apt-get install -y --no-install-recommends libfabric-dev wrk libaio-dev numactl && \ apt-get install -y --no-install-recommends libfabric-dev wrk libaio-dev numactl && \

28
docker/llm/serving/xpu/docker/oneccl-binding.patch
View file

@ -1,14 +1,32 @@
diff --git a/src/gpu/dpcpp_ccl.cpp b/src/gpu/dpcpp_ccl.cpp diff --git a/src/gpu/dpcpp_ccl.cpp b/src/gpu/dpcpp_ccl.cpp
index 3bd8087..c5b5ce3 100644 index bb1f236..289d490 100644
--- a/src/gpu/dpcpp_ccl.cpp --- a/src/gpu/dpcpp_ccl.cpp
+++ b/src/gpu/dpcpp_ccl.cpp +++ b/src/gpu/dpcpp_ccl.cpp
@@ -689,7 +689,8 @@ c10::intrusive_ptr<ProcessGroupCCL::AsyncWorkCCL> XPUCCLStubs::allreduce_(std::v @@ -80,6 +80,16 @@ int get_sync_only(int init_value = 0) {
return tmp_sync_only;
}
+char* get_cached_env(const char* var_name) {
+ static char* cached_value = nullptr;
+ if (!cached_value) {
+ cached_value = std::getenv(var_name);
+ if (!cached_value) {
+ cached_value = "0";
+ }
+ }
+ return cached_value;
+}
#define CCL_KERNEL_SUBMIT(cmd, q) \
({bool profile_barrier = (xpu::is_profiler_enabled()); \
@@ -759,6 +769,10 @@ c10::intrusive_ptr<ProcessGroupCCL::AsyncWorkCCL> XPUCCLStubs::allreduce_impl(st
stream, stream,
attr), stream.get_native()); attr), stream.get_native());
}); });
- // printf("Use One CCL allreduce.\n"); + const char* env_value = get_cached_env("IPEX_LLM_CCL_ENABLE_NATIVE_WAIT");
+ if (env_value && std::string(env_value) == "1") {
+ stream.get_native().wait(); + stream.get_native().wait();
+ // printf("Use One CCL allreduce.\n"); + }
// printf("Use One CCL allreduce.\n");
return ret_evt; return ret_evt;
}, },
c10d::OpType::ALLREDUCE);