[Mlir-commits] [mlir] [MLIR][AMDGPU] Add OCP FP8 support for new hardware (PR #106160)

[Paul C Fuqua]

================
@@ -47,6 +47,13 @@ struct Chipset {
   DEFINE_COMP_OPERATOR(>)
   DEFINE_COMP_OPERATOR(>=)
 #undef DEFINE_COMP_OPERATOR
+
+  bool isGfx940() const {
+    return majorVersion == 9 && minorVersion >= 4 && minorVersion < 5;
----------------
pcf000 wrote:

Faulty merge.  It made more sense before steppingVersion, as >=0x40 && < 0x50.

https://github.com/llvm/llvm-project/pull/106160