[PATCH] [x86] Implement a faster vector population count based on the PSHUFB in-register LUT technique.
Owen Anderson
resistor at mac.com
Fri May 29 09:53:32 PDT 2015
> On May 29, 2015, at 8:20 AM, Bruno Cardoso Lopes <bruno.cardoso at gmail.com> wrote:
>
> My patch to improve vector legalization for pop count from http://reviews.llvm.org/D10002 <http://reviews.llvm.org/D10002> is certainly a win here, but won't certainly beat using ARM64's native popcnt on vXi8 and building the results for wider types on top of that!
I wasn’t actually suggesting not using ARM64’s native vXi8 instructions, but rather the rest of the sequence that synthesizes wider lane pop counts on top of it.
—Owen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150529/b7538b27/attachment.html>
More information about the llvm-commits
mailing list