mirror of
https://github.com/deepseek-ai/DeepGEMM.git
synced 2026-11-05 00:47:26 +00:00
[PR #376] Fix NVRTC compilation for HyperConnection kernels #398
Labels
No labels
pull-request
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
deepseek-ai/DeepGEMM#398
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/deepseek-ai/DeepGEMM/pull/376
Author: @nya-a-cat
Created: 7/11/2026
Status: 🔄 Open
Base:
main← Head:agent/fix-nvrtc-hc-pdl-header📝 Commits (1)
7ac7449fix NVRTC HyperConnection PDL declaration📊 Changes
2 files changed (+4 additions, -0 deletions)
View changed files
📝
deep_gemm/include/deep_gemm/impls/sm100_tf32_hc_prenorm_gemm.cuh(+2 -0)📝
deep_gemm/include/deep_gemm/impls/sm90_tf32_hc_prenorm_gemm.cuh(+2 -0)📄 Description
Summary
<cuda_device_runtime_api.h>in the SM90 and SM100 HyperConnection pre-norm GEMM kernels.cudaGridDependencySynchronize()declaration available to the NVRTC compilation path.Root cause
Both HyperConnection kernels call the Programmatic Dependent Launch synchronization intrinsic. With
DG_JIT_USE_NVRTC=1, the current include chain does not expose its device-runtime declaration, so NVRTC fails with:NVCC provides the declaration implicitly in this path, which is why the failure is NVRTC-specific.
Impact
This restores NVRTC compilation for both SM90 and SM100 HyperConnection kernels. It does not change kernel logic, synchronization behavior, or launch configuration.
Validation
git diff --check origin/main...HEADGPU runtime validation was not performed because the Actions runner did not have an NVIDIA GPU.
Fixes #359
🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.