[Mlir-commits] [mlir] Added free-threading CPython mode support in MLIR Python bindings (PR #107103)

[Stella Laurenzo]

stellaraccident wrote:

> I'll also +1 that adding fine-grain locking may be problematic. The MLIR Python builders are already slow, and adding lots of fine grained locking, even uncontended locking, will make that worse. But in JAX, as in Stella's case, we never access the same IR objects from multiple threads, and if we did need to do that we could add client-side locking. I think that's by far the most important mode to support.

Off topic, but one of these days, I'm going to have to create a stripped down, build only (probably nanobind based) fast version of the API. Not at the top of the list right now, but yeah... slow. On the torch side, we only use the non-ODS API so that we have more flexibility here in the future.

https://github.com/llvm/llvm-project/pull/107103