This revision was automatically updated to reflect the committed changes.
Closed by commit rG9e7c9967c3fd: Additionally set f32 mode with
denormal-fp-math (authored by dcandler).
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D122589/new/
https://revi
arsenm accepted this revision.
arsenm added a comment.
This revision is now accepted and ready to land.
LGTM. I played around with this and it seems to have the overriding behavior I
would expect
Comment at: clang/test/CodeGen/denormalfpmode-f32.c:21
+
+// CHECK-LABEL: main
+
dcandler added a comment.
The issue I found was trying to use getDefaultDenormalModeForType during
constant folding to account for denormals (https://reviews.llvm.org/D116952).
Setting denormal-fp-math to a non-IEEE mode without specifying
denormal-fp-math-f32 still results in the denormal-fp-m
arsenm added a comment.
I think my intent here was to not emit denormal-fp-math-f32 at all if it wasn't
specified (i.e. denormal-fp-math-f32 is a special override only really used for
AMDGPU)
The interaction between the two and the default is supposed to be handled by
the backend interpretatio
dcandler updated this revision to Diff 418865.
dcandler added a comment.
Added a test that checks attributes based on the -fdenormal-fp-math and
-fdenormal-fp-math-f32 flags.
Only the cases where -fdenormal-fp-math is set to preserve-sign or
positive-zero and -fdenormal-fp-math-f32 is unset are
arsenm added a comment.
Testcase?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D122589/new/
https://reviews.llvm.org/D122589
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org
dcandler created this revision.
dcandler added reviewers: arsenm, spatel.
Herald added a project: All.
dcandler requested review of this revision.
Herald added subscribers: cfe-commits, MaskRay, wdng.
Herald added a project: clang.
When the denormal-fp-math option is used, this should set the
deno
