janwas 27539d25e8 improvements to memcpy; cleaned up the whole thing and ia32 is now done (knock on wood) ...

executive summary: memcpy2 aka. ia32_memcpy is a drop-in replacement
that is much faster than memcpy (7% on small buffers, 300% on 192K
buffers)

This was SVN commit r2729.

2005-09-15 00:51:59 +00:00

..

unix

Linux/GCC Compat, added some includes, ported inline assembly stuff to gnu syntax (and misc. cleanup in ia32.cpp)

2005-09-12 23:37:52 +00:00

win

read intel doc on multicore, realized it'd break our #cpu detect (which is actually critical because it determines safety of TSC). wrote new CPU detect code which should fix it (wasn't easy, grr); cannot test for lack of Pentium EE hardware ;p

2005-09-14 16:58:10 +00:00

cpu.h

read intel doc on multicore, realized it'd break our #cpu detect (which is actually critical because it determines safety of TSC). wrote new CPU detect code which should fix it (wasn't easy, grr); cannot test for lack of Pentium EE hardware ;p

2005-09-14 16:58:10 +00:00

dir_watch.h

…

gfx.h

…

ia32.asm

improvements to memcpy; cleaned up the whole thing and ia32 is now done (knock on wood)

2005-09-15 00:51:59 +00:00

ia32.cpp

improvements to memcpy; cleaned up the whole thing and ia32 is now done (knock on wood)

2005-09-15 00:51:59 +00:00

ia32.h

improvements to memcpy; cleaned up the whole thing and ia32 is now done (knock on wood)

2005-09-15 00:51:59 +00:00

os.h

…

snd.h

…

sysdep.cpp

read intel doc on multicore, realized it'd break our #cpu detect (which is actually critical because it determines safety of TSC). wrote new CPU detect code which should fix it (wasn't easy, grr); cannot test for lack of Pentium EE hardware ;p

2005-09-14 16:58:10 +00:00

sysdep.h

read intel doc on multicore, realized it'd break our #cpu detect (which is actually critical because it determines safety of TSC). wrote new CPU detect code which should fix it (wasn't easy, grr); cannot test for lack of Pentium EE hardware ;p

2005-09-14 16:58:10 +00:00