[Mlir-commits] [mlir] [mlir] GEMM Hopper Tensor Core Integration Test (PR #81478)

[Guray Ozen]

grypp wrote:

*[`mlir/test/Integration/GPU/CUDA/sm90/python/tools/matmulBuilder.py` line 177 at r3](https://reviewable.io/reviews/llvm/llvm-project/81478#-Ns3Xwhd6TXc8uSESvck:-Ns3tCnL2SBZvaW7LWvO:byaf2qb) ([raw file](https://github.com/llvm/llvm-project/blob/6da82f445e66be30e31674aaf9aa7f075ac02735/mlir/test/Integration/GPU/CUDA/sm90/python/tools/matmulBuilder.py#L177)):*
<details><summary><i>Previously, jpienaar (Jacques Pienaar) wrote…</i></summary><blockquote>

So none of the nvgpu types have python bindings yet, and just parse the string variant?
</blockquote></details>

We can add that. for the time being we are parsing strings 


<!-- Sent from Reviewable.io -->


https://github.com/llvm/llvm-project/pull/81478