microsoft/mscclpp
1
0
Fork 0
mirror of https://github.com/microsoft/mscclpp.git synced 2026-04-19 14:29:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

Torch integration (#692)

Browse Source 
Reorganize current native algorithm implementation and DSL algorithm
implementation.
Provide unified API for DSL algo and native algo and provide interface
to tune the algo
Provide interface for pytorch integration with native API and DSL

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com>
Co-authored-by: chhwang <8018170+chhwang@users.noreply.github.com>
This commit is contained in:
Binyang Li
2026-01-21 20:32:24 -08:00
committed by GitHub
parent 78ce9fac8d
commit a707273701
156 changed files with 6107 additions and 4076 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
docs/tutorials/python-api.md
View File

@@ -15,9 +15,9 @@ from mscclpp import (
Transport,
)
from mscclpp.utils import GpuBuffer
import mscclpp.comm as mscclpp_comm
def create_connection(group: mscclpp_comm.CommGroup, transport: str):
def create_connection(group: mscclpp.CommGroup, transport: str):
remote_nghrs = list(range(group.nranks))
remote_nghrs.remove(group.my_rank)
if transport == "NVLink":
@@ -30,7 +30,7 @@ def create_connection(group: mscclpp_comm.CommGroup, transport: str):
return connections
if __name__ == "__main__":
mscclpp_group = mscclpp_comm.CommGroup(MPI.COMM_WORLD)
mscclpp_group = mscclpp.CommGroup(MPI.COMM_WORLD)
connections = create_connection(mscclpp_group, "NVLink")
nelems = 1024
memory = GpuBuffer(nelem, dtype=cp.int32)