[X86] AMD Zen 3: AVX XMM moves are zero-cycle

I've verified this with llvm-exegesis. This is not limited to zero registers.
2021-05-07 16:23:38 +03:00 · 2021-05-07 16:23:38 +03:00 · ee020b930d
parent 9db4203883
commit ee020b930d
2 changed files with 1246 additions and 1220 deletions
--- a/llvm/lib/Target/X86/X86ScheduleZnver3.td
+++ b/llvm/lib/Target/X86/X86ScheduleZnver3.td
@ -1474,10 +1474,12 @@ def : IsOptimizableRegisterMove<[
    // SSE variants.
    MOVAPSrr, MOVUPSrr,
    MOVAPDrr, MOVUPDrr,
-    MOVDQArr, MOVDQUrr
+    MOVDQArr, MOVDQUrr,

    // AVX variants.
-    // FIXME
+    VMOVAPSrr, VMOVUPSrr,
+    VMOVAPDrr, VMOVUPDrr,
+    VMOVDQArr, VMOVDQUrr
  ], TruePred >
 ]>;

--- a/llvm/test/tools/llvm-mca/X86/Znver3/reg-move-elimination-avx-xmm.s
+++ b/llvm/test/tools/llvm-mca/X86/Znver3/reg-move-elimination-avx-xmm.s