2010/2/26 Roman Tetelman <[email protected]>: > On 02/25/2010 02:40 PM, Michael Menegakis wrote: >> >> I noticed that it is slower than x86 and I suspect it is because >> unlike vm_x86, it does not directly write opcodes but first makes an >> assembly array of lines that in turn get assembled. >> >> It does certainly make a more beautiful code. >> >> However, If it's the reason final loading is noticeably slower, it >> might have been a sane decision by id software to go directly with >> opcodes writing. >> _______________________________________________ >> ioquake3 mailing list >> [email protected] >> http://lists.ioquake.org/listinfo.cgi/ioquake3-ioquake.org >> By sending this message I agree to love ioquake3 and libsdl. >> > > this was done to make writing the code easier, and so far, no one has > volunteered to update it to use opcodes directly like vm_x86.c
It certainly was easier to direct it to win64 that way; since microsoft made the decision of using a completely different calling convention than unix on win64, calling had to be written differently for it. But since mingw-w64 had the __attribute__ ((sysv_abi)) feature of gcc, it was easier to simply declared the functions confined in the vm code with that attribute. _______________________________________________ ioquake3 mailing list [email protected] http://lists.ioquake.org/listinfo.cgi/ioquake3-ioquake.org By sending this message I agree to love ioquake3 and libsdl.
