mirror of
https://github.com/kvcache-ai/Mooncake.git
synced 2026-11-05 01:17:55 +00:00
[PR #973] [MERGED] Support AMDGPU (refactor CUDA-alike) #1931
Labels
No labels
ascend
auto-closed
bug
bug
CI/Build
Common
documentation
enhancement
feature request
good first issue
help wanted
high priority
MetaX
nightly-failure
pull-request
PyTorch Backend
question
RFC
Roadmap
stale
Store
Tests
Transfer Engine
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
kvcache-ai/Mooncake#1931
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
📋 Pull Request Information
Original PR: https://github.com/kvcache-ai/Mooncake/pull/973
Author: @yeahdongcn
Created: 10/27/2025
Status: ✅ Merged
Merged: 10/31/2025
Merged by: @stmatengss
Base:
main← Head:xd/cuda_alike📝 Commits (6)
f1c75cfSupport AMDGPU (refactor CUDA alike)93ea477Address review comments8c3490cFix build5334d5aAddress review comments524bce8Minor update502c203Minor update📊 Changes
15 files changed (+210 additions, -232 deletions)
View changed files
📝
mooncake-common/common.cmake(+22 -2)📝
mooncake-transfer-engine/example/transfer_engine_bench.cpp(+7 -23)📝
mooncake-transfer-engine/example/transfer_engine_bench_with_notify.cpp(+8 -24)📝
mooncake-transfer-engine/example/transfer_engine_bench_with_retry.cpp(+7 -23)📝
mooncake-transfer-engine/example/transfer_engine_validator.cpp(+10 -24)➕
mooncake-transfer-engine/include/cuda_alike.h(+15 -0)➕
mooncake-transfer-engine/include/gpu_vendor/hip.h(+70 -0)📝
mooncake-transfer-engine/include/gpu_vendor/musa.h(+41 -42)📝
mooncake-transfer-engine/src/CMakeLists.txt(+5 -0)📝
mooncake-transfer-engine/src/memory_location.cpp(+3 -15)📝
mooncake-transfer-engine/src/topology.cpp(+3 -19)📝
mooncake-transfer-engine/src/transport/rdma_transport/rdma_context.cpp(+1 -8)📝
mooncake-transfer-engine/src/transport/tcp_transport/tcp_transport.cpp(+7 -14)📝
mooncake-transfer-engine/tests/rdma_transport_test.cpp(+7 -25)📝
mooncake-transfer-engine/tests/tcp_transport_test.cpp(+4 -13)📄 Description
This PR mainly adds support for AMDGPU and refactors the CUDA-alike implementations.
Testing Done
🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.