Changes between Version 3 and Version 4 of LibCSSE
- Timestamp:
- May 22, 2014, 4:33:19 PM (12 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
LibCSSE
v3 v4 15 15 }}} 16 16 17 ||= Idea =||= Westmere =||= Sandy Bridge =||= Ivy Bridge =||= Penryn =||17 ||= Idea =||= Westmere =||= Sandy Bridge =||= Ivy Bridge =||= Penryn =|| 18 18 || Replace `dec` with `sub` || none || none || none || || 19 19 || Use movsd instead of movsq || slightly slower || slightly slower || 6% faster || || … … 35 35 Now testing the overlap case: 36 36 37 ||= Idea =||= Westmere =||= Sandy Bridge =||= Ivy Bridge =||= Penryn =||37 ||= Idea =||= Westmere =||= Sandy Bridge =||= Ivy Bridge =||= Penryn =|| 38 38 || `movaps` 64 at a time || 56% faster || 56% faster || 56% faster || 48% faster || 39 39 || Above using leaq || 50% faster || 56% faster || 60% faster || 52% faster ||
