Wishbringer
05-27-2002, 03:17 AM
Only a question of understanding..
libEQ.a is compiled by GCC? (think so, read someting about that, so i think it's C-Code, not assembler or similar..)
With GCC 3.0.x there are options to configure target machine (-march=ix86).
Would there be runtime-improvements to have an i686 version on P2/P3-Systems?
What's is current target of libEQ.a? i386?
Or with new GCC 3.1 there are options to make code for P3/P4s, Athlons and AthlonXPs..
e.g. -march=pentium3 -msse -mfpmath=sse
-march=athlon-xp -m3dnow -mfpmath=sse
I can't test it for myself (hehe, don't have sourcecode of libEQ.a --> and even don't want it :p ).
But perhaps owners of sourcecode could test if there were some runtime-improvements, and perhaps you could release some multiple version...
Of course not all possiblities of compiler options, but maybe 3 or 4 versions...
As i said, only a suggestion, not really necessary....
libEQ.a is compiled by GCC? (think so, read someting about that, so i think it's C-Code, not assembler or similar..)
With GCC 3.0.x there are options to configure target machine (-march=ix86).
Would there be runtime-improvements to have an i686 version on P2/P3-Systems?
What's is current target of libEQ.a? i386?
Or with new GCC 3.1 there are options to make code for P3/P4s, Athlons and AthlonXPs..
e.g. -march=pentium3 -msse -mfpmath=sse
-march=athlon-xp -m3dnow -mfpmath=sse
I can't test it for myself (hehe, don't have sourcecode of libEQ.a --> and even don't want it :p ).
But perhaps owners of sourcecode could test if there were some runtime-improvements, and perhaps you could release some multiple version...
Of course not all possiblities of compiler options, but maybe 3 or 4 versions...
As i said, only a suggestion, not really necessary....