Tried this on an athlon-xp 2000+. Works. This is what 'cat /proc/cpuinfo' says:
processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 6 model name : AMD Athlon(tm) XP 2000+ stepping : 2 cpu MHz : 1666.378 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow bogomips : 3339.11 Hope this will be fixed soon. > -----Ursprüngliche Nachricht----- > Von: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Im Auftrag > von Ondrej Hošek > Gesendet: Samstag, 6. Oktober 2007 19:17 > An: [email protected] > Betreff: Re: [hlds_linux] Team Fortress 2/Dedicated Server updated > > Both Intel and AMD manuals say that if CPUID returns flags > saying that "FPU" and "CMOV" are present, then "FCMOV" is present too. > > If you want to be sure, just try it the hard way. Copy the > following code into a text file (named "fcmov.s" in this example): > > .code32 > .text > .global _start > _start: > fldpi > fldpi > mov $0x0,%eax > cmp $0x0,%eax > fcmove %st(1),%st > mov $0x1,%eax > int $0x80 > > Then run this: > > as -o fcmov.o fcmov.s > ld -o fcmov fcmov.o > ./fcmov > > If nothing happens, your processor can do FCMOV. If you get a > SIGILL (illegal instruction exception), congratulations, your > processor is FCMOV-less. > > You'll have to have some sort of assembler with AT&T syntax > support (normally called "as" or "gas") as well as a linker > ("ld") installed on your machine. Alternatively, you can > assemble and link it on a different machine with the same > major and minor kernel version, such as 2.6, and then run it > on your target. > > (The code runs only on Linux (the last two statements, asking > the system to quit the process, are Linux-specific). It > shouldn't touch any of your devices, unless the kernel is > defective, so there shouldn't be any data loss, but just like > everything originating from me, no warranty is given.) > > If you still get bone weight errors even though this program > completes without breaking, then the culprit isn't FCMOV. > > ~~ Ondra > > On 06.10.07 16:54 Uhr, Steve Sumichrast wrote: > > I Don't think this is accurate information -- My linux box > (AMD Barton 3000+ running Arch Linux kernel 2.6.20) is having > the bad bone weight error -- and I have both of those listed > in the output. > > > > However, I do not have FCMOV -- I have CMOV support.. > Guessing that's what's doing it. > > > > -----Original Message----- > > From: [EMAIL PROTECTED] > > [mailto:[EMAIL PROTECTED] On Behalf > Of Ondrej > > Hošek > > Sent: Saturday, October 06, 2007 08:17 AM > > To: [email protected] > > Subject: Re: [hlds_linux] Team Fortress 2/Dedicated Server updated > > > > grep ext /proc/cpuinfo > > > > If the output includes "fpu" and "cmov", chances are your > machine can do it. > > > > As for the "why"... Windows and Linux use different compilers. GCC > > probably compiles into code using FCMOV while cl.exe doesn't. > > > > ~~ Ondra > > > > On 06.10.07 5:59 Uhr, Kevin J. Anderson wrote: > > > >> In what processor range(s) was this FCMOV introduced? > >> > >> Alfred Reynolds wrote: > >> > >>> Your CPU needs to support the FCMOV instruction, if you > get the bone > >>> weight problem then chances are it doesn't. > >>> > >>> - Alfred > >>> > >>> Daron Dodd wrote: > >>> > >>>> i see no fix to the bad bone weight crashes > >>>> > >>>> On 10/5/07, Jason Ruymen <[EMAIL PROTECTED]> wrote: > >>>> > >>>>> Required updates to Team Fortress have been released. > Please run > >>>>> hldsupdatetool to receive them. The specific changes include: > >>>>> > >>>>> - Fixed certain crashes with paged pool memory > >>>>> - Added a warning for when paged pool memory is low > >>>>> - Fixed some prediction errors with player avoidance > >>>>> - Fixed columns and row bunching up in server browser > >>>>> - Fixed a rare bug where the wrong launcher.dll was > being used to > >>>>> launch the game > >>>>> - Fixed teleporter/spawn doorway exploit > >>>>> > >>>>> Jason > >>>>> > > _______________________________________________ > To unsubscribe, edit your list preferences, or view the list > archives, please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > > No virus found in this incoming message. > Checked by AVG Free Edition. > Version: 7.5.488 / Virus Database: 269.14.2/1052 - Release > Date: 05.10.2007 18:53 > > No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.488 / Virus Database: 269.14.3/1054 - Release Date: 06.10.2007 19:12 _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
