Re: [Dri-devel] Some bsd-3-0-0-branch benchmarks.
On Thu, 2002-06-27 at 15:46, Keith Whitwell wrote: > So, instead of DRM_OS_COPYFROMUSR_NC, maybe DRM_COPY_FROM_USER_UNCHECKED > might be clearer. > > Similarly, DRM_OS_KRNFROMUSR is pretty cryptic -- maybe > DRM_COPY_FROM_USER_IOCTL or something? > > Oh, and I just found DRM_OS_FETCHU_32_NC -- that's ugly... I > > How about: > > DRM_OS_COPYFROMUSR_NC > -- DRM_COPY_FROM_USER_UNCHECKED > DRM_OS_COPYFROMUSR-- DRM_COPY_FROM_USER > DRM_OS_KRNFROMUSR -- DRM_COPY_FROM_USER_IOCTL > DRM_OS_FETCHU_32_NC -- DRM_GET_USER_UNCHECKED > > and so on. Yeah, I thought the _NCs were ugly when I named them. I did that plus DRM_OS_IOCTL -> DRM_IOCTL_ARGS, DRM_OS_DELAY->DRM_OS_UDELAY, and killed DRM_OS_RETURN (not used in the os-independent code anyway). Do you know what's up with the i810 code in linux? It's complaining about not enough args to do_munmap_Rsmp_8fdbbff9. It would be nice to complete the whole drm building on Linux. -- Eric Anholt <[EMAIL PROTECTED]> http://gladstone.uoregon.edu/~eanholt/dri/ --- This sf.net email is sponsored by:ThinkGeek Bringing you mounds of caffeinated joy. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] Some bsd-3-0-0-branch benchmarks.
Keith Whitwell wrote: > Eric Anholt wrote: > >> Well, I've got most of the FreeBSD troubles straightened out I think. I >> went ahead and did some glxgears benchmarks, waiting for the numbers to >> stabilize, of gentoo vs freebsd-current. >> >> System is a 128MB 2xCeleron517 (BP6, OCed), diskless, booting gentoo or >> -current off of a -current system. Video at 1280x1024x60hz, 16 bit. >> >> gentoo 1.1, kernel 2.4.19-gentoo-r5 (custom, p2 opts but nothing special >> think) >> bsd-3-0-0-branch as of a few days ago >> kernel modules from bsd-3-0-0-branch from today >> twm >> >> FreeBSD-current as of a few days ago. >> WITNESS and INVARIANTS disabled. >> bsd-3-0-0-branch as of a few days ago >> kernel modules from bsd-3-0-0-branch from today. >> No wm (sshing in). >> >> Radeon 64MB VIVO, no pageflip: >> linux: 1325 fps >> bsd: 1324 >> >> Rage 128 Pro: >> linux: 581 >> bsd: 582 >> >> Matrox G400 >> linux: 923 >> bsd: 755 >> >> Only the Matrox had problems, don't know what that was.Still, I'm >> very excited about the Radeon numbers. > > > Nice. > > I'm just looking a little at the code, wonder if there can be some name > changes to simplify the macros a little. > > We're using DRM_ prefixes for os-abstractions already, like DRM_DEBUG, > so I don't think it's necessary to further specialize the namespace to > DRM_OS_ ... (doh) So, instead of DRM_OS_COPYFROMUSR_NC, maybe DRM_COPY_FROM_USER_UNCHECKED might be clearer. Similarly, DRM_OS_KRNFROMUSR is pretty cryptic -- maybe DRM_COPY_FROM_USER_IOCTL or something? Oh, and I just found DRM_OS_FETCHU_32_NC -- that's ugly... I How about: DRM_OS_COPYFROMUSR_NC -- DRM_COPY_FROM_USER_UNCHECKED DRM_OS_COPYFROMUSR -- DRM_COPY_FROM_USER DRM_OS_KRNFROMUSR -- DRM_COPY_FROM_USER_IOCTL DRM_OS_FETCHU_32_NC -- DRM_GET_USER_UNCHECKED and so on. Keith --- This sf.net email is sponsored by:ThinkGeek Bringing you mounds of caffeinated joy. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] Some bsd-3-0-0-branch benchmarks.
Eric Anholt wrote: > Well, I've got most of the FreeBSD troubles straightened out I think. I > went ahead and did some glxgears benchmarks, waiting for the numbers to > stabilize, of gentoo vs freebsd-current. > > System is a 128MB 2xCeleron517 (BP6, OCed), diskless, booting gentoo or > -current off of a -current system. Video at 1280x1024x60hz, 16 bit. > > gentoo 1.1, kernel 2.4.19-gentoo-r5 (custom, p2 opts but nothing special > think) > bsd-3-0-0-branch as of a few days ago > kernel modules from bsd-3-0-0-branch from today > twm > > FreeBSD-current as of a few days ago. > WITNESS and INVARIANTS disabled. > bsd-3-0-0-branch as of a few days ago > kernel modules from bsd-3-0-0-branch from today. > No wm (sshing in). > > Radeon 64MB VIVO, no pageflip: > linux: 1325 fps > bsd: 1324 > > Rage 128 Pro: > linux: 581 > bsd: 582 > > Matrox G400 > linux: 923 > bsd: 755 > > Only the Matrox had problems, don't know what that was.Still, I'm > very excited about the Radeon numbers. Nice. I'm just looking a little at the code, wonder if there can be some name changes to simplify the macros a little. We're using DRM_ prefixes for os-abstractions already, like DRM_DEBUG, so I don't think it's necessary to further specialize the namespace to DRM_OS_ --- This sf.net email is sponsored by:ThinkGeek Bringing you mounds of caffeinated joy. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
[Dri-devel] Some bsd-3-0-0-branch benchmarks.
Well, I've got most of the FreeBSD troubles straightened out I think. I went ahead and did some glxgears benchmarks, waiting for the numbers to stabilize, of gentoo vs freebsd-current. System is a 128MB 2xCeleron517 (BP6, OCed), diskless, booting gentoo or -current off of a -current system. Video at 1280x1024x60hz, 16 bit. gentoo 1.1, kernel 2.4.19-gentoo-r5 (custom, p2 opts but nothing special think) bsd-3-0-0-branch as of a few days ago kernel modules from bsd-3-0-0-branch from today twm FreeBSD-current as of a few days ago. WITNESS and INVARIANTS disabled. bsd-3-0-0-branch as of a few days ago kernel modules from bsd-3-0-0-branch from today. No wm (sshing in). Radeon 64MB VIVO, no pageflip: linux: 1325 fps bsd: 1324 Rage 128 Pro: linux: 581 bsd: 582 Matrox G400 linux: 923 bsd: 755 Only the Matrox had problems, don't know what that was.Still, I'm very excited about the Radeon numbers. -- Eric Anholt <[EMAIL PROTECTED]> http://gladstone.uoregon.edu/~eanholt/dri/ --- Sponsored by: ThinkGeek at http://www.ThinkGeek.com/ ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel