[fpc-pascal] Crypt++ 5.5.2 Tiger::Transform converted toDCPCrypt2.Compress, AV: Compiler Optimization Bug with SSE2 ? Or just programming bug ?HELP NEEDED :) ?!
skybuck2000 at hotmail.com
Thu Mar 20 22:54:16 CET 2008
> FPC needs to preserve ebx, esi and edi.
So it's push ebx + push esi + push edi vs pushad. And same for pop.
According to my AMD optimization manual.
pushad requires 6 latency.
push register requires 3 latency.
So pushad should be faster.
More information about the fpc-pascal