[PATCH] D71604: [X86] Add strict fma support

[Craig Topper via Phabricator via llvm-commits]

craig.topper added inline comments.


================
Comment at: llvm/lib/Target/X86/X86InstrFragmentsSIMD.td:498
+def X86strict_fma: SDNode<"ISD::STRICT_FMA",   SDTFPTernaryOp, [SDNPCommutative, SDNPHasChain]>;
+def X86Fmadd     : PatFrags<(ops node:$src1, node:$src2, node:$src3),
+                            [(X86strict_fma node:$src1, node:$src2, node:$src3),
----------------
I'd prefer we call this X86any_fmadd and update the users. That will make things more consistent.


================
Comment at: llvm/test/CodeGen/X86/fp-strict-scalar.ll:585
+define double @fma_f64(double %a, double %b, double %c) nounwind strictfp {
+; SSE-X86-LABEL: fma_f64:
+; SSE-X86:       # %bb.0:
----------------
The AVX1 check lines seem to be missing? Use the update script


================
Comment at: llvm/test/CodeGen/X86/vec-strict-128.ll:223
+; SSE-LABEL: f13:
+; SSE:       # %bb.0:
+; SSE:         call{{[l|q]}} fmaf
----------------
Use the script


================
Comment at: llvm/test/CodeGen/X86/vec-strict-256.ll:160
+; SSE-LABEL: f13:
+; SSE:       # %bb.0:
+; SSE:         call{{[l|q]}} fmaf
----------------
Use the script


================
Comment at: llvm/test/CodeGen/X86/vec-strict-512.ll:156
 
+define <16 x float> @f13(<16 x float> %a, <16 x float> %b, <16 x float> %c) #0 {
+; SSE-LABEL: f13:
----------------
This was not generated with the update_llc_test_checks.py script, regenerate it.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D71604/new/

https://reviews.llvm.org/D71604