[Mlir-commits] [mlir] [mlir][linalg] Support subtracting accumulation in partial reduction … (PR #214033)

[llvmlistbot at llvm.org]

pstarkcdpr wrote:

Oh, I see @FedericoBruzzone. You can tell that I'm not at all familiar with the LLVM community.

Before making my previous example, I tried doing some (AI assisted) research, and it seem to agree with your question (1) above. In a `linalg.generic`, the order is not guaranteed, so any overflow flags _ought_ to work regardless of how you permute the inputs. I assumed this to be true when trying to construct an example of why things would go wrong anyways.

>  I think that we should know whether the current tiling mechanism is currently blocked by these flags.
Not sure what you mean by blocked. Do you mean that we should try to figure out if there's an example with the existing ops (like `addi`) could produce incorrect results due to the current flag propagation or if subtract is possibly making things different?

https://github.com/llvm/llvm-project/pull/214033