[Mlir-commits] [mlir] [MLIR][Linalg] Remove matmul_transpose variants (PR #147961)

[llvmlistbot at llvm.org]

MaheshRavishankar wrote:

> > Could we get some time to manage the churn with these changes?
> 
> Of course!! No rush from my side. Let me know if you need some help.

It would be ideal if we had a deprecation path instead of just a switch. The changes needed here are
- Torch-MLIR (and all other "front end dialects" like StableHLO, etc.) will need to be updated to not generate the batch-matmul operations
- and downstream compilers (like IREE) need to make sure that they can navigate away from this.

In LLVM this would be something that would be deprecated on one release and removed on the next release (it is pretty much a break the world change).  I dont think all the front end dialect folks are paying attention to this.

https://github.com/llvm/llvm-project/pull/147961