[PATCH] D71293: AMDGPU: Generate the correct sequence of code for FDIV32 when correctly-rounded-divide-sqrt is set

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp:568-569
+    // is enabled.
+    if (!HasCorrectlyRoundedDivideSqrt)
+      return false;
+    SafeFast = false;
----------------
cfang wrote:
> arsenm wrote:
> > The attribute should not be considered at all. Only the fpmath metadata matters. If -cl-fp32-correctly-rounded-divide-sqrt is specified, the regular fdiv instruction should behave correctly.
> Do you mean that in AMDGPUCodeGenPrepare, we should check the fpmath metadata to keep regular fdiv (instead of an intrinsic) when  -cl-fp32-correctly-rounded-divide-sqrt is specified?
> 
> The issue is, when -cl-fp32-correctly-rounded-divide-sqrt is specified, a simple v_rcp is generated for a fdiv. Apparently the codegen produces the wrong sequence of code for a "regular" fdiv. 
> 
Yes. We shouldn't even have an IR attribute for this flag. The interpretation of the flag is entirely represented in the use of the !fpmath metadata. 


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D71293/new/

https://reviews.llvm.org/D71293