[Mlir-commits] [mlir] [MLIR][XeGPU] Introduce `xegpu::uArch` usage in target-sensitive passes (PR #163801)

[Jianhui Li]

================
@@ -43,7 +43,12 @@ def XeGPUPropagateLayout : Pass<"xegpu-propagate-layout"> {
   let options = [Option<
     "printOnly", "print-analysis-only", "bool",
     /*default=*/"false",
-    "Print the result of layout propagation analysis and exit.">];
+    "Print the result of layout propagation analysis and exit.">,
+    Option<
+    "assumeUnrolled", "assume-unrolled", "bool",
----------------
Jianhui-Li wrote:

Can this option be an enumeration, so the propagation could be applied to "lane", "inst", and "subgroup" parameters? High-level implies lower level will be propagated, so "assumeUnrolled = true" can be replaced to "level = lane" here and the options are more extensible. 

https://github.com/llvm/llvm-project/pull/163801