Skip to content

Enable hipBLASLt GEMM for gfx115x#5082

Merged
causten merged 1 commit into
ROCm:developfrom
jomohamm:johny/enable-hipblaslt-gfx115x
Jul 24, 2026
Merged

Enable hipBLASLt GEMM for gfx115x#5082
causten merged 1 commit into
ROCm:developfrom
jomohamm:johny/enable-hipblaslt-gfx115x

Conversation

@jomohamm

Copy link
Copy Markdown
Contributor

Summary

Adds gfx115 to hipblaslt_supported_impl() so gfx1150/1151/1152/1153 (Strix Halo/Point, RDNA3.5) use the hipBLASLt GEMM path.

Previously hipblaslt_supported_impl() gated hipBLASLt on gfx110x and gfx120x but not gfx115x, so on these parts the lowering pass fell back to rocBLAS for every GEMM. rocBLAS lacks BF16/INT8 Tensile solutions for these parts, so quantized dot ops returned rocblas_status_not_implemented at runtime while FP16/FP32 worked. hipBLASLt ships tuned BF16/INT8 kernels for the gfx115 family, so this allows it to use the hipBLASLt path.

Change

  • src/targets/gpu/device_name.cpp: add starts_with(gfx_name, "gfx115") to the hipBLASLt-supported predicate.

Test plan

  • Build with MIGRAPHX_USE_HIPBLASLT enabled and run quantized (BF16/INT8) dot ops on a gfx115x device; confirm they no longer fall back to rocBLAS / return rocblas_status_not_implemented.
  • Confirm FP16/FP32 GEMM behavior unchanged on gfx115x.

Made with Cursor

Add support for gfx115x so gfx1150/1151/1152/1153 (Strix Halo/Point,
RDNA3.5) are supported in hipBLASLt.

hipblaslt_supported_impl() gated hipBLASLt on gfx110x and gfx120x but not
gfx115x, so on these parts the lowering pass fell back to rocBLAS for
every GEMM. rocBLAS lacks BF16/INT8 Tensile solutions for these parts, so
quantized dot ops returned rocblas_status_not_implemented at runtime while
FP16/FP32 worked. hipBLASLt ships tuned BF16/INT8 kernels for the gfx115
family, so allow it to use the hipBLASLt path.

Signed-off-by: Johny Shaik Mohammad <[email protected]>
@jomohamm
jomohamm requested a review from causten as a code owner July 21, 2026 05:35
@github-actions

Copy link
Copy Markdown
Contributor

Thank you for your contribution! Since this is an external pull request, a maintainer must review PR and add the "ok-to-test" label if it is approved for testing.

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Enables the hipBLASLt GEMM path on RDNA3.5 (gfx115x) GPUs by extending the device-family gate used by the GPU backend, preventing fallback to rocBLAS for GEMM on those parts.

Changes:

  • Extend hipblaslt_supported_impl() to treat gfx115* as hipBLASLt-supported (alongside gfx110x/gfx120x and select gfx94x/gfx95x).

Comment on lines 142 to +145
return (gfx_name == "gfx90a" or (starts_with(gfx_name, "gfx94") and gfx_name >= "gfx942") or
(starts_with(gfx_name, "gfx95") and gfx_name >= "gfx950") or
starts_with(gfx_name, "gfx110") or starts_with(gfx_name, "gfx120"));
starts_with(gfx_name, "gfx110") or starts_with(gfx_name, "gfx115") or
starts_with(gfx_name, "gfx120"));
@jomohamm
jomohamm requested a review from kahmed10 July 22, 2026 05:59

@CharlieL7 CharlieL7 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@causten
causten merged commit c488cd1 into ROCm:develop Jul 24, 2026
39 of 41 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants