[all-commits] [llvm/llvm-project] ed0ea8: [NVGPU] Add FP8 (e4m3/e5m2) support to nvgpu.mma.s...

[WMC via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: ed0ea836f6cc029b3c8ec9b2308f5e94038dac03
      https://github.com/llvm/llvm-project/commit/ed0ea836f6cc029b3c8ec9b2308f5e94038dac03
  Author: WMC <tnwilly at gmail.com>
  Date:   2026-07-16 (Thu, 16 Jul 2026)

  Changed paths:
    M mlir/lib/Conversion/NVGPUToNVVM/NVGPUToNVVM.cpp
    M mlir/lib/Dialect/NVGPU/IR/NVGPUDialect.cpp
    A mlir/test/Conversion/NVGPUToNVVM/nvgpu-to-nvvm-mma-fp8.mlir

  Log Message:
  -----------
  [NVGPU] Add FP8 (e4m3/e5m2) support to nvgpu.mma.sync (#207342)

This PR is stacked on #207307 

## Description 

Add FP8 E4M3/E5M2 support for `nvgpu.mma.sync`, building on the
NVVM-level
support in #207307. This adds NVGPU verification and lowering support,
registers
the supported `m16n8k16`/`m16n8k32` MMA variants, and adds coverage for
NVGPU
lowering and NVVM verification.

## Test

```sh
mlir-opt mlir/test/Conversion/NVGPUToNVVM/nvgpu-to-nvvm-fp8.mlir \
  -convert-nvgpu-to-nvvm -split-input-file
```

Co-authored with Claude Sonnet 5

---------

Signed-off-by: weimin023 <tnwilly at gmail.com>



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications