[Piglit] [PATCH 1/1] cl: Fix tolerance for SP division

[Tom Stellard]

On Fri, Jan 23, 2015 at 06:51:39PM -0500, Jan Vesely wrote:
> The specs say 2.5 ulp, but we only take integer values
> 

Pushed, thanks!

> Signed-off-by: Jan Vesely <jan.vesely at rutgers.edu>
> ---
> 
> R600: Fail -> Pass
> 
>  tests/cl/program/execute/vector-arithmetic-float4.program_test | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/tests/cl/program/execute/vector-arithmetic-float4.program_test b/tests/cl/program/execute/vector-arithmetic-float4.program_test
> index 7127bb9..a80a973 100644
> --- a/tests/cl/program/execute/vector-arithmetic-float4.program_test
> +++ b/tests/cl/program/execute/vector-arithmetic-float4.program_test
> @@ -91,21 +91,21 @@ name: pos div pos, neg div neg, pos div neg, neg div pos
>  kernel_name: div
>  arg_in:  1 float4             8.5 -21.25 11.25  -21
>  arg_in:  2 float4            4.25     -5    -3 5.25
> -arg_out: 0 buffer float4[1]     2   4.25 -3.75   -4
> +arg_out: 0 buffer float4[1]     2   4.25 -3.75   -4 tolerance 2
>  
>  [test]
>  name: 0 div pos, 0 div neg, pos div 0, neg div 0
>  kernel_name: div
>  arg_in:  1 float4               0  0 11.25  -21
>  arg_in:  2 float4            4.25 -5     0    0
> -arg_out: 0 buffer float4[1]     0  0   inf -inf
> +arg_out: 0 buffer float4[1]     0  0   inf -inf tolerance 2
>  
>  [test]
>  name: 0 div inf, inf div 0, inf div inf, nan div 0
>  kernel_name: div
>  arg_in:  1 float4               0  inf inf nan
>  arg_in:  2 float4             inf    0 inf   0
> -arg_out: 0 buffer float4[1]     0  inf nan nan
> +arg_out: 0 buffer float4[1]     0  inf nan nan tolerance 2
>  
>  ## Unary plus ##
>  
> -- 
> 2.1.0
> 
> _______________________________________________
> Piglit mailing list
> Piglit at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/piglit