microsoft/mscclpp
1
0
Fork 0
mirror of https://github.com/microsoft/mscclpp.git synced 2026-04-19 22:39:11 +00:00
Code Issues Packages Projects Releases Wiki Activity

Use GpuIpcMem for NVLS connections (#719)

Browse Source 
* Now `NvlsConnection` internally reuses `GpuIpcMem` for multicast
memory handling.
* Removed unnecessary barriers from `connectNvlsCollective()` (CUDA API
handles this automatically).
* Updated `GpuIpcMem::map()` and `GpuIpcMem::mapMulticast()` to return a
shared pointer with custom deleter for unmapping, which prevents misuse
of raw pointers and reduces states to be stored in the `GpuIpcMem`
instance.
* Now for `RuntimeIpc` type handles, for consistency with other types,
`cudaIpcOpenMemHandle` will be called in `GpuIpcMem::map()` instead of
the ctor of `GpuIpcMem`.

---------

Co-authored-by: Binyang Li <binyli@microsoft.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com>
Co-authored-by: Binyang2014 <9415966+Binyang2014@users.noreply.github.com>
This commit is contained in:
Changho Hwang
2026-01-14 21:16:04 -08:00
committed by GitHub
parent c2a87302bd
commit 105239fc6c
7 changed files with 226 additions and 290 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
python/csrc/switch_channel_py.cpp
View File

@@ -29,8 +29,7 @@ void register_nvls(nb::module_& m) {
});
nb::class_<NvlsConnection>(m, "NvlsConnection")
.def("bind_allocated_memory", &NvlsConnection::bindAllocatedMemory, nb::arg("device_ptr"), nb::arg("size"))
.def("get_multicast_min_granularity", &NvlsConnection::getMultiCastMinGranularity);
.def("bind_allocated_memory", &NvlsConnection::bindAllocatedMemory, nb::arg("device_ptr"), nb::arg("size"));
m.def("connect_nvls_collective", &connectNvlsCollective, nb::arg("communicator"), nb::arg("all_ranks"),
nb::arg("buffer_size"));