[llvm-commits] [llvm] r58875 - /llvm/trunk/test/CodeGen/PowerPC/fnegsel.ll

[Chris Lattner]

On Nov 7, 2008, at 4:40 PM, Dale Johannesen wrote:

> Author: johannes
> Date: Fri Nov  7 18:40:24 2008
> New Revision: 58875
>
> URL: http://llvm.org/viewvc/llvm-project?rev=58875&view=rev
> Log:
> Xfail an incorrect test.

Hi Dale,

Please just remove this and move the potential optimization  
opportunity to lib/Target/PowerPC/README.txt.  Thanks!

-Chris

>
>
>
> Modified:
>    llvm/trunk/test/CodeGen/PowerPC/fnegsel.ll
>
> Modified: llvm/trunk/test/CodeGen/PowerPC/fnegsel.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/PowerPC/fnegsel.ll?rev=58875&r1=58874&r2=58875&view=diff
>
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> ======================================================================
> --- llvm/trunk/test/CodeGen/PowerPC/fnegsel.ll (original)
> +++ llvm/trunk/test/CodeGen/PowerPC/fnegsel.ll Fri Nov  7 18:40:24  
> 2008
> @@ -1,5 +1,12 @@
> ; RUN: llvm-as < %s | llc -march=ppc32 | not grep fneg
>
> +; FSEL is not IEEE-safe, and should not be generated except with
> +; -enable-finite-only-fp-math.  Further, it can't be generated for
> +; GT ever; it tests for GE.  Leaving in the test for now as it may
> +; be useful as a basis for a correct test.
> +;
> +; XFAIL: *
> +
> define double @test_FNEG_sel(double %A, double %B, double %C) {
>         %D = sub double -0.000000e+00, %A               ; <double>  
> [#uses=1]
>         %Cond = fcmp ugt double %D, -0.000000e+00               ;  
> <i1> [#uses=1]
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits