microsoft/mscclpp v0.9.0

MSCCL++ v0.9.0

Repository: microsoft/mscclpp

Tag: v0.9.0

Published: 2026-04-09T01:15:40Z

Prerelease: no

Release notes:

What's Changed

• Fix lint.sh by @chhwang in https://github.com/microsoft/mscclpp/pull/652
• Update the port channel tutorial doc by @chhwang in https://github.com/microsoft/mscclpp/pull/653
• Fix test script by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/655
• Update EndpointConfig interfaces by @chhwang in https://github.com/microsoft/mscclpp/pull/651
• New allreduce algo for small message size by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/647
• Fix docs by @chhwang in https://github.com/microsoft/mscclpp/pull/656
• Fix docs version by @chhwang in https://github.com/microsoft/mscclpp/pull/659
• Exclude irrelevant files from workflow triggers by @chhwang in https://github.com/microsoft/mscclpp/pull/663
• Improving DSL documentation by @caiomcbr in https://github.com/microsoft/mscclpp/pull/650
• Add exclude paths under pipeline triggers by @chhwang in https://github.com/microsoft/mscclpp/pull/664
• Test peer accessibility after deployment by @chhwang in https://github.com/microsoft/mscclpp/pull/661
• Rename nvls* files by @chhwang in https://github.com/microsoft/mscclpp/pull/660
• Fix #651 by @chhwang in https://github.com/microsoft/mscclpp/pull/662
• Add token pool for cuCreate API by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/628
• Auto-detect CUDA arch in CMake GPU check by @chhwang in https://github.com/microsoft/mscclpp/pull/666
• FP8 support for Allreduce by @seagater in https://github.com/microsoft/mscclpp/pull/646
• Resolve IBVerbs Loading Issues by @caiomcbr in https://github.com/microsoft/mscclpp/pull/648
• Fixes for no-IB systems by @chhwang in https://github.com/microsoft/mscclpp/pull/667
• Integrate MSCCL++ DSL to torch workload by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/620
• Add a new logger by @chhwang in https://github.com/microsoft/mscclpp/pull/668
• upgrade codeql to v3 by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/676
• IB stack enhancements & bug fixes by @chhwang in https://github.com/microsoft/mscclpp/pull/673
• Support Synchronous Initialization for Proxy Service by @caiomcbr in https://github.com/microsoft/mscclpp/pull/679
• Supporting New Packet Kernel Operation at Executor by @caiomcbr in https://github.com/microsoft/mscclpp/pull/677
• connect() APIs changed to return an instance instead of a shared_ptr by @chhwang in https://github.com/microsoft/mscclpp/pull/680
• Fix Minor Issue Proxy Python Interface by @caiomcbr in https://github.com/microsoft/mscclpp/pull/685
• Revise the mscclpp datatype by @seagater in https://github.com/microsoft/mscclpp/pull/671
• Fix Error in Non IB Env at Executor by @caiomcbr in https://github.com/microsoft/mscclpp/pull/686
• No IB Env CI Test by @caiomcbr in https://github.com/microsoft/mscclpp/pull/687
• Fix Python bindings and tests by @chhwang in https://github.com/microsoft/mscclpp/pull/690
• DSL Quick Start by @caiomcbr in https://github.com/microsoft/mscclpp/pull/689
• Add CudaDeviceGuard by @chhwang in https://github.com/microsoft/mscclpp/pull/691
• Optimized logger by @chhwang in https://github.com/microsoft/mscclpp/pull/693
• Build fixes by @chhwang in https://github.com/microsoft/mscclpp/pull/696
• Add an IB multi-node tutorial by @chhwang in https://github.com/microsoft/mscclpp/pull/702
• Creating Documentation Section for MSCCL++ DSL by @caiomcbr in https://github.com/microsoft/mscclpp/pull/706
• Make IB more configurable by @chhwang in https://github.com/microsoft/mscclpp/pull/703
• Improve DSL Documentation by @caiomcbr in https://github.com/microsoft/mscclpp/pull/707
• Add handle cache for AMD platform by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/698
• Add copilot-instructions.md by @chhwang in https://github.com/microsoft/mscclpp/pull/602
• Use uncached memory on Rocm platform to avoid hang by @qishilu in https://github.com/microsoft/mscclpp/pull/711
• Replace __HIP_PLATFORM_AMD__ to use internal macro by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/712
• Rename P2P log subsys into GPU by @chhwang in https://github.com/microsoft/mscclpp/pull/716
• Minor fixes by @chhwang in https://github.com/microsoft/mscclpp/pull/715
• Remove UB std:: declarations by @chhwang in https://github.com/microsoft/mscclpp/pull/709
• Tune the nThreadsPerBlock for FP8 and Half datatype on MI300 by @seagater in https://github.com/microsoft/mscclpp/pull/694
• Update container images for pipeline by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/717
• Add CUDA 13.0 Docker images by @chhwang in https://github.com/microsoft/mscclpp/pull/720
• Bypassing SSCA alerts by @chhwang in https://github.com/microsoft/mscclpp/pull/721
• Add GpuIpcMemHandle by @chhwang in https://github.com/microsoft/mscclpp/pull/704
• Reduce CI build time by @chhwang in https://github.com/microsoft/mscclpp/pull/723
• Use GpuIpcMem for NVLS connections by @chhwang in https://github.com/microsoft/mscclpp/pull/719
• Fix ci issue by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/727
• Fix ci pipeline failure by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/729
• Torch integration by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/692
• fp8 nvls support (e5m2 and e4m3) by @mahdiehghazim in https://github.com/microsoft/mscclpp/pull/730
• Support versioning for mscclpp document by @seagater in https://github.com/microsoft/mscclpp/pull/724
• Revert "Support versioning for mscclpp document (#724)" by @seagater in https://github.com/microsoft/mscclpp/pull/734
• Use native GPU architecture when NVIDIA GPU is detected; otherwise fall back to multi-arch build. by @mahdiehghazim in https://github.com/microsoft/mscclpp/pull/732
• Update document versioning for PR #724 by @seagater in https://github.com/microsoft/mscclpp/pull/735
• Fix the relative path extraction on github page by @seagater in https://github.com/microsoft/mscclpp/pull/739
• Support multi-node in MemoryChannel tutorial by @chhwang in https://github.com/microsoft/mscclpp/pull/726
• Address comments for PR #692 by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/733
• Refactor reduce kernel by @Binyang2014 in https://github.com/microsoft/mscclpp/pull/738
• Fix cpplint error in main branch by @seagater in https://github.com/microsoft/mscclpp/pull/740
• Update copilot-instructions.md by @chhwang in…