Phoronix IRC Log: 2009-03-14
bkero: http://staff.osuosl.org/~bkero/newmonitor.jpg :) yays
bkero: Gonna have to drill another monitor stand to get the other monitor to rotate.
maligor: maybe the pcie bench will now work since they released stream sdk 1.4beta
Kano: hi maligor
Kano: still playing with the ati bench?
maligor: never got any results from it
Kano: a benchmark without result ist funny
maligor: I wonder why the stream sdk comes in a weird .run and a rpm inside it
Kano: to confuse you
maligor: and they changed the name from amdstream to atistream
Kano: they dont know what they want
Kano: maybe they want to sell ati
maligor: but they now have glorious support for integer ops in brook!
Kano: it seems the ssl test is the only benchmark a phenom 2 wins against intel
Ivanovic: michaellarabel: you made mistake in your news entry about "VDPAU + OpenGL 3.0 On Gallium3D This Summer?"
Ivanovic: that is: so far it is not sure if x.org will be in GSOC or not
Ivanovic: they applied, sure, but the list of accepted orgs won't be out before wednesday
michaellarabel: Ivanovic: I am assuming it will be, it's such an important project and it has been in every year so far.
Ivanovic: (okay, chances are high, that they will be in, but it is not guranteed)
Ivanovic: you should probably add that it is not guranteed so far that they will be in or just make it "they applied for Summer of Code with proposed projects about..."
Ivanovic: and i hope wesnoth will make it again, too...
Kano: i guess a Q9300 would need 3900 mhz to beat a phenem ii@ 3ghz with openssl, something must be really wrong there.
maligor: very interesting definitely
maligor: maybe it trashes the cache constantly
maligor: phenoms would have less latency to ram
Kano: can somebody with a fast intel retest the 2nd?
Kano: the cache is small in q9300,maybe thats what hurts
maligor: tried a icc compiled openssl? (if it's possible)
maligor: it could be a gccism
Kano: but for fairness the p should use the same compiler then too
redeeman: i doubt it, i believe openssl has x86_
redeeman: 64 assembler
maligor: but time to go play with embedded bootloaders
redeeman: which one?
Kano: well he used gcc 4.3.3, lenny has 4.3.2
Kano: but if that makes the diff
redeeman: it doesn't
Kano: i used pts 1.6 to compare it better, just forgot for the first test the LANG= override
Kano: therefore the value looks a bit weird
maligor: das u-boot
Kano: michaellarabel: did you find the issue with fps?
michaellarabel: Kano: Haven't had any test system on yet
Kano: michaellarabel: it was even with a the very basic fgl_glxgears test
Kano: no test runs that shows fps?
Kano: michaellarabel: missing deps for Triangle Slammer, a) correct debian name for the one which is shown: libopengl-perl, still missing: SDL/App.pm , which is in libsdl-perl
Kano: but the benchmark still shows nothing in the window...
Kano: Possible Package Names: libglew, libglew1.5-dev, glew-devel
Kano: the first can not be correct for that depends
Kano: when you need headers
Kano: michaellarabel: i think it has something to do with your browser detection
Kano: not fully sure, but i ran some benchmarks with returned
Kano: when it does not return i see with htop that pts runs @ 100% with one core
Kano: michaellarabel: how to enable debugging what is called AFTER the test run
Kano: i thought a 3450 is slow,but intel q35 is extra slow with nexuiz
Deanjo: Kano, intel Core's have always been weak in openSSL when compared to the AMD's
Kano: but why?
Kano: Deanjo: btw your display driver is outdated ;)
Deanjo: No idea, not sure if the openSSL has some very AMD specific optimization or not.
Deanjo: Dammit, go out drinking for one night and NV releases a new driver when I'm not around.
Deanjo: I should have known, it was friday, Battlestar Galactica, Numb3rs, and nv driver update :P
Kano: dont forget wine update, of couse since yesterday in kanotix
Deanjo: Hmmm, no changelog yet though
Kano: no, didnt find a log too
Deanjo: Rats, Aaron isn't in chat either
Kano: a cuda error in the board, but i dont see much about it
Deanjo: Ya not sure about that, I'll have to see if I can replicate it later this weekend
Deanjo: brb, got to put this driver into play
Deanjo: Well I see that a few sli bugs have been fixed
michaellarabel: ugh, ASUS Eee Top with Linux is bad.
maligor: Eee Top?
maligor: touchscreen model huh
maligor: michaellarabel, if you have one.. can you see what touchscreen hw it has?
maligor: oh, it's a desktop system with a touchscreen
maligor: with a mobility radeon, heh
michaellarabel: maligor: Its a TouchPack Resistive Touch Screen
michaellarabel: WiFi only works on 9.04. Touchscreen isn't working with 8.10 or 9.04. And the xf86-video-intel driver breaks with it (currently working on that bug)
maligor: oh.. the specs I see mention it having a mobility radeon 3450
michaellarabel: that is for the ET1603
michaellarabel: this is the ET1602
maligor: do they ship it with linux option?
maligor: the device concept is definately very interesting
michaellarabel: No, only with Windows
michaellarabel: With the -intel driver I only get the backlight on
dandel: any one have a few hints on doing a git disect?
Deanjo: michaellarabel, one other thing that was fixed on the new nvidia drivers is that "SLI" "auto" works once again.
borkdox: tried the catalyst 9.4 RC yet?
borkdox: anybody here*
dandel: and i can't use it ><; they dropped the card i use
borkdox: yea, at least open source drivers are gettin better
borkdox: hopefully with gallium 3d will be faster than mesa
dandel: i wonder how long it'll be before they decide to cut my card from the windows driver.
dandel: my desktop uses an Radeon x1900GT
Kano: hi michaellarabel , i saw you try to autoinstall deps with sudo on ubuntu, you could do the same with debian live, sudo is without pass there
Kano: in most cases the package names are the same
dandel: hey kano.
Kano: hi dandel
dandel: can ya help me with a quick git-bisect so i can help fix a bug which is knocking out my acpi?
Kano: well i only used git bisect with wine
Kano: i followed the wine instructions, using ccache as compiler and so
dandel: i did 4 versions of kernel dmesg to help pinpoint the error :/
dandel: 2.6.26 when they integrated start_secondary is when the bug first appeared.
dandel: 2.6.29rc7 had the error change tho.
Kano: [ 17.856117] toshiba_acpi: Unknown parameter `hotkeys_over_acpi'
Kano: thats your error?
Kano: grep hotkeys_over_acpi /etc/modprobe.d/*
Kano: show me this
dandel: options toshiba_acpi hotkeys_over_acpi=1
Kano: what file
dandel: how do i figure that out?
dandel: i did the command... cat /etc/modprobe.d/* | grep hotkeys_over_acpi
dandel: oh... found it... it's under... toshiba_acpi.modprobe
Kano: why dont you use my command
Kano: that would have shown the filename
dandel: /etc/modprobe.d/toshiba_acpi.modprobe:options toshiba_acpi hotkeys_over_acpi=1
dandel: the last thing i pasted is the output of your command.
Kano: dandel: delete that file or put a # in front of that line
dandel: 1 sec
dandel: i'll go ahead and make that modification, however it's worth noting that toshiba_acpi is not a module i have open after the system fully booted
dandel: and this is on the kernel which has acpi working
Kano: be sure that the editor you use does not write backupfiles
Kano: use the grep again after the mod
dandel: quick request, do i add ath5k to the blacklist if i don't want it to load at boot?
dandel: changed to two files
Kano: echo blacklist ath5k > /etc/modprobe/ath5k
Kano: for example
dandel: one with a ~ at the end, the other has the modification you requested.
Kano: rm all backup files!
dandel: i doubt this will fix it tho ><;
dandel: reboots laptop
dandel: which kernel should i check on... 2.6.29rc7, 2.6.26, or 2.6.27
dandel: keep in mind all of those are broken ones.
dandel: it disabled acpi_irq again
Kano: the error can not exist anymnore
dandel: 2.6.29rc7 gives a mini crash dump in the dmesg
dandel: search that page for acpi_irq
dandel: the crash dump lines start at... [ 2.073273]  __report_bad_irq+0x27/0x90
dandel: a separate dump occurs in this one... http://launchpadlibrarian.net/23767544/dmesg-2.6.26-020626-generic.log
dandel: and it started at this line... [ 4.119859]  __report_bad_irq+0x27/0x90
Kano: did you try irqpoll
dandel: i'll check when i update it.
dandel: pulls 2.6.29rc8 to check again.
dandel: kernel.maps_protect what's that?
Kano: dont know
dandel: irqpoll still gives disabled irq
dandel: i did try... acpi=noirq and that worked, but it causes issues elsewhere.
Kano: maybe blacklist the toshiba module
dandel: no, mp-bios bug is the other error if i set acpi to not use an irq
dandel: my special function keys are what breaks when i set acpi=noirq
Kano: didyou ever blacklist the toshiba_acpi ?
dandel: no fix
dandel: my issue is surrounding the start_secondary integration and threading of the acpi checks i think.
dandel: 0o interesting
dandel: i just found it.
borkdox: catalyst 9.4 rc1 is looking for me, but its having trouble reaching me....
redeeman: Creative Sound Blaster X-Fi 5.1 VX XtremeAudio XtremeMusic SW Titanium Gamer XMOD Fatal1ty Pro Champion Series Go! FPS
Lightkey: no "Live!" or "AWE64"? :-)