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.
* Added cl_khr_fp16 extension support for commonfns test (issue #142, commonfns)
* Added missing header due to presubmit check
* Corrected radians/degrees ulp calculations + cosmetic fixes
* Corrected presubmit code format
* Corrections related to code review
* Moved string format helper to test_common in separate header
* Added clang format for last commit
* Corrections related to code review
* Modified mix verification procedure for half type to only report max error
* Removed redundant condition for logging mix verification
* Corrected generator limits for half tests
* Unified common functions tests due to preparation for adding cl_khr_fp16 support
* Renamed base structure, few cosmetic corrections
* Added corrections due to code review
* Removed comment separators
* Added review related corrections
