Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
Description
Currently, alignment of allocation is fixed with 64 and this enables most of the architectures, but not all L1D prefetching systems are the same and some of the architectures are using double prefetching like x86_64. Include a matrix of alignment values to extend the cache alignments.
Attachments
Issue Links
- links to