[PATCH] D153024: AMDGPU: Correctly lower llvm.exp2.f32

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/docs/AMDGPUUsage.rst:976
+                                             half). Not implemented for double. Hardware provides
+                                             1ULP accuracy for float, and 0.51ULP for half. Float
+                                             instruction does not natively support denormal
----------------
rampitec wrote:
> Typo 0.51ULP?
If it's a typo it's a typo in the ISA manual


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D153024/new/

https://reviews.llvm.org/D153024