================
@@ -1,24 +1,24 @@
-! RUN: %flang_fc1 -ffast-math -emit-mlir -o - %s | FileCheck %s
--check-prefixes=CHECK%if target=x86_64{{.*}} %{,CHECK-KIND10%}%if
flang-supports-f128-math %{,CHECK-KIND16%}
+! RUN: %flang_fc1 -ffast-real-mod -emit-mlir -o - %s | FileCheck %s
--check-prefixes=CHECK-FRM%if target=x86_64{{.*}} %{,CHECK-FRM-KIND10%}%if
flang-supports-f128-math %{,CHECK-FRM-KIND16%}
+! RUN: %flang_fc1 -ffast-math -emit-mlir -o - %s | FileCheck %s
--check-prefixes=CHECK-FRM%if target=x86_64{{.*}} %{,CHECK-FRM-KIND10%}%if
flang-supports-f128-math %{,CHECK-FRM-KIND16%}
! RUN: %flang_fc1 -ffast-math -fno-fast-real-mod -emit-mlir -o - %s |
FileCheck %s --check-prefixes=CHECK-NFRM%if target=x86_64{{.*}}
%{,CHECK-NFRM-KIND10%}%if flang-supports-f128-math %{,CHECK-NFRM-KIND16%}
----------------
mjklemm wrote:
Since `-fno-fast-real-mod` is sort of the default, so that part should already
be tested via `test/Lower/Intrinsics/mod.f90`. I guess, I could add a fourth
test that ensures that the result of `-fno-fast-real-mod` matches the one of
the `mod.f90` test. Would that be useful?
https://github.com/llvm/llvm-project/pull/167118
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
