[Mlir-commits] [mlir] [mlir][vector] Update docs for scalable vectors (PR #101842)

Andrzej WarzyƄski llvmlistbot at llvm.org
Mon Aug 5 02:34:36 PDT 2024


================
@@ -286,10 +299,11 @@ Consider a vector of rank n with static sizes `{s_0, ... s_{n-1}}` (i.e. an MLIR
 `vector<s_0x...s_{n-1}xf32>`). Lowering such an `n-D` MLIR vector type to an
 LLVM descriptor can be done by either:
 
-1.  Flattening to a `1-D` vector: `!llvm<"(s_0*...*s_{n-1})xfloat">` in the MLIR
+1.  Nested aggregate type of `1-D` vector:
----------------
banach-space wrote:

If you scroll further down, you will see that Option 1 and Option 2 are discussed in the reverse order. I just wanted to improve the flow (also got confused myself when reading this). 

I can move this change to a separate patch, but am concerned that that's just extra noise.

https://github.com/llvm/llvm-project/pull/101842


More information about the Mlir-commits mailing list