I've commited code to speed up the performance of multi-precision integer multiplication on processors which support SSE2 instructions (i.e. Intel Pentium 4).
The speedup is about 2x on my test platform, a Pentium 4 Mobile 1700.
To enable these on Linux/Un*x, use the --enable-sse2 switch with the configure command.
For the win32 platform, see file Makefile.mak - you have to add switch /DOPTIMIZE_SSE2 to ASFLAGS.
Log in to post a comment.