You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
CpuMath Enhancement: Preamble for hardware intrinsics implementation #830
while (!aligned) { do scalar operation; } // preamble
Do vectorized operation using ReadAligned
while (!end) { do scalar operation; }
For large arrays, especially those that cross cache line or page boundaries, doing this should save some measurable amount of time.
changed the title [-]Preamble for hardware intrinsics implementation[/-][+]CpuMath Enhancement: Preamble for hardware intrinsics implementation[/+]on Sep 6, 2018
Activity
[-]Preamble for hardware intrinsics implementation[/-][+]CpuMath Enhancement: Preamble for hardware intrinsics implementation[/+]