systems-assistant[bot]
f05be9efb3
* Initial refactoring work, including using build targets, and settable MSCCLPP_ROOT, MSCCLPP_SOURCE, MSCCLPP_APPLY_PATCHES. * Another large refactor of MSCCLPP cmake to make all portions targets with appropriate dependencies. This should include all paths to the final target: starting with a full mscclpp install, starting with custom mscclpp and/or json source code, or from submodules + optional patches. * Update whitespace Findmscclpp_nccl_static.cmake --------- Co-authored-by: Corey Derochie <corey.derochie@amd.com> Co-authored-by: corey-derochie-amd <161367113+corey-derochie-amd@users.noreply.github.com>
52 γραμμές
2.2 KiB
Diff
52 γραμμές
2.2 KiB
Diff
diff --git a/CMakeLists.txt b/CMakeLists.txt
|
|
index a95a8e5..62b4f22 100644
|
|
--- a/CMakeLists.txt
|
|
+++ b/CMakeLists.txt
|
|
@@ -96,6 +96,24 @@ include(${PROJECT_SOURCE_DIR}/cmake/AddFormatTargets.cmake)
|
|
|
|
# Find ibverbs and libnuma
|
|
find_package(IBVerbs)
|
|
+
|
|
+# Check if IBV_ACCESS_RELAXED_ORDERING exists in infiniband/verbs.h
|
|
+# Disable use of this symbol in mscclpp/src/ib.cc if it does not exist
|
|
+if(IBVERBS_FOUND)
|
|
+ try_compile(HAS_IBV_ACCESS_RELAXED_ORDERING
|
|
+ ${CMAKE_BINARY_DIR}
|
|
+ "${EXT_SOURCE}/check_ibv_access_relaxed_ordering.cc"
|
|
+ CMAKE_FLAGS "-DINCLUDE_DIRECTORIES=${IBVERBS_INCLUDE_DIRS}"
|
|
+ OUTPUT_VARIABLE try_compile_output
|
|
+ )
|
|
+ message(STATUS "try_compile_output: ${try_compile_output}")
|
|
+ if(NOT HAS_IBV_ACCESS_RELAXED_ORDERING)
|
|
+ message(WARNING "IBV_ACCESS_RELAXED_ORDERING does not exist in ${IBVERBS_INCLUDE_DIRS}/infiniband/verbs.h. Disabling this symbol in mscclpp/src/ib.cc.")
|
|
+ else()
|
|
+ message(STATUS "IBV_ACCESS_RELAXED_ORDERING exists in ${IBVERBS_INCLUDE_DIRS}/infiniband/verbs.h.")
|
|
+ endif()
|
|
+endif()
|
|
+
|
|
find_package(NUMA REQUIRED)
|
|
find_package(Threads REQUIRED)
|
|
|
|
diff --git a/src/ib.cc b/src/ib.cc
|
|
index d9d72d1..bddd4a8 100644
|
|
--- a/src/ib.cc
|
|
+++ b/src/ib.cc
|
|
@@ -48,9 +48,17 @@ IbMr::IbMr(ibv_pd* pd, void* buff, std::size_t size) : buff(buff) {
|
|
}
|
|
uintptr_t addr = reinterpret_cast<uintptr_t>(buff) & -pageSize;
|
|
std::size_t pages = (size + (reinterpret_cast<uintptr_t>(buff) - addr) + pageSize - 1) / pageSize;
|
|
+
|
|
+#if defined(HAS_IBV_ACCESS_RELAXED_ORDERING)
|
|
this->mr = IBVerbs::ibv_reg_mr2(pd, reinterpret_cast<void*>(addr), pages * pageSize,
|
|
IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_REMOTE_WRITE | IBV_ACCESS_REMOTE_READ |
|
|
IBV_ACCESS_RELAXED_ORDERING | IBV_ACCESS_REMOTE_ATOMIC);
|
|
+#else
|
|
+ this->mr = IBVerbs::ibv_reg_mr2(pd, reinterpret_cast<void*>(addr), pages * pageSize,
|
|
+ IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_REMOTE_WRITE | IBV_ACCESS_REMOTE_READ |
|
|
+ IBV_ACCESS_REMOTE_ATOMIC);
|
|
+#endif
|
|
+
|
|
if (this->mr == nullptr) {
|
|
std::stringstream err;
|
|
err << "ibv_reg_mr failed (errno " << errno << ")";
|