Ahmed b4c3bf2af2 Fixes for basic explicit_s2v and commonfns degrees for cl_half (#2024) ...

Basic explicit_s2v:

The verification step was always using round to even when converting a
float to half even for round to zero cores.

Commonfns degrees:

The verification step was only taking into account infinities and not
values that over/underflow. This resulted in an incorrect error
calculation. E.g:

double cpu_result = 175668.85998711039;
cl_half gpu_result = 31743; // this is 65504 when converting to float,
we overflowed.
float error = (cpu_result - gpu_result) * some_factor;

The fix adds the check if( (cl_half) reference == test ) before
calculating the error.

2024-08-13 09:52:07 -07:00

16 KiB

Raw Blame History

View Raw