SETI MB CUDA for Linux

[riofl]

I don't understand something. The developers obviously know that cuda devices hate vlar and vhar workunits. why can't they put the smarts that are in the perl script directly into boinc so that as it downloads the workunit it reads the angles etc and decides where to assign it at that point? it should be an easy thing to implement saving lots of trouble for people like me who have a card that locks up at the slightest hint of a vlar and having to run something external to make sure 'proper' workunits are fed to cuda?

seems to me that this ia a gross oversight leaving this out of boinc.

[Josef W. Segur]

I don't understand something. The developers obviously know that cuda devices hate vlar and vhar workunits. why can't they put the smarts that are in the perl script directly into boinc so that as it downloads the workunit it reads the angles etc and decides where to assign it at that point? it should be an easy thing to implement saving lots of trouble for people like me who have a card that locks up at the slightest hint of a vlar and having to run something external to make sure 'proper' workunits are fed to cuda?

seems to me that this ia a gross oversight leaving this out of boinc.

True, the S@H CUDA app has it's greatest advantage for midrange work. The stock BOINC won't be changed as you suggest for the simple reason that it's designed to support common features of use to all projects. BOINC is separate from S@H even though they both derive from Classic S@H in part. OTOH, it's LGPL open source so a special version could be built, feel free to volunteer for that work. 

As far as the S@H project providing some additional features, I figure once volunteers begin donating enough money to resolve ongoing problems the staff might be able to find time to work something out. Meanwhile, many of those who have CUDA crunching turned off while using the computer for other things probably don't even notice the difficulty of VLAR. There isn't really any similar issue with VHAR, the main thing there is simply the number of tasks which have to be downloaded when most are VHAR 'shorties'.
                                                                                     Joe

[riofl]

makes sense. i keep forgetting about  other projects since i only do seti. maybe a shell after the list download would be in order to run a script if it was there. i always find it difficult trapsing around others' code trying to figure out what is what since a majority of the time it is not well commented if at all. i have never looked at boinc source. maybe once they have a stable version that works with multiple devices and properly reports them i might dig in and see what is what. i doubt i would submit mods for this because it would create a separate seti branch of boinc which probably would not be wise. of course, a simple cmdline flag --seti or some such could turn the extra code on or off.. hmm... have to think abit about this.

[lordvader]

Hey all.

I recently got a computation error with one of my CUDA WU's, but it wasn't a segfault. Maybe there's some info in there for you guys :

http://setiathome.berkeley.edu/result.php?resultid=1347109626

[Raistmer]

Hey all.

I recently got a computation error with one of my CUDA WU's, but it wasn't a segfault. Maybe there's some info in there for you guys :

http://setiathome.berkeley.edu/result.php?resultid=1347109626

Thanks, but it's known error.
It has even its own sticky thread on main ("-12"). It's some limitation of current CUDA code.
But sometimes such error could appear when card go in broken state.
So if you see many of such errors, then it's equal of many "-9 overflows" that no validated versus wingman => host should be rebooted.
Unfortunately, I have one such buggy card so had chance to see this "-12" for dozens of results at once (usually it's pretty rare error).

[letni]

Hey folks, I'm trying to set up 2 separate CUDA devices on a dedicated 64 bit Linux system with the setiathome-6.08.CUDA_2.2_x86_64-pc-linux-gnu.  I have the system running with 1 CUDA (9600GSO) no problem, but I decided to stick in the 8800GTS from my desktop machine and now I'm getting this message:

CUDA device: GeForce 8800GTS (driver version 0, compute capability 1.0, 320MB, est. 41GFLOPS)
CUDA device (not used): GeForce 8800GTS (driver version 0, compute capability 1.0, 320MB, est. 41GFLOPS)

For some reason It doesn't detect my 9600 GSO as an 9600 anymore, but still uses it (the card is physically hot).  The 8800 I added in as a secondary card is cool to the tough and I don't see a thread running for it.   Here is the LSPCI output:

02:00.0 VGA compatible controller: nVidia Corporation GeForce 9600 GSO (rev a2)
03:00.0 VGA compatible controller: nVidia Corporation G80 [GeForce 8800 GTS] (rev a2)

Is it even possible to use CUDA with two different devices?  If so, what am I missing?

Thanks,

Letni

[sunu]

After a small hiatus I'm back 

now i dont know about desktop settings much but there is one change i made in the past few weeks with nvidia-settings. i unchecked Sync to VBlank in xvideo settings and also unchecked sync to vblank and allow flipping in the opengl settings. wasnt sure what they did but there seemed to be no difference.   should they be checked?

Sync to vblank and flipping don't have anything to do with cuda computations. They are about tearing.

power mizer which seems to not have settings says adaptive clocking enabled performance level2 perforamce mode desktop. level2 is the 3d settings above however i remember when i first got the card, performance mode said maximum performance and somewhere along the line it changed to desktop. since the other settings are the same i can only assume it is a function of which driver is being used for which text shows up.

I don't like powermizer at all, but it doesn't seem you have a problem there.

since 6.9.0 reports 2 teslas, could it be possible it is mixing up which device is 0 and which is 1? because it is completely odd since the tesla is running gpu 500mhz and memory 900mhz so it should be considerably slower. it rates both devices it thinks are teslas at 74gflops yet the 285 is rated by 6.6.11 as 127gflops

Again, what boinc reports is irrelevant, just cosmetic. What cuda client sees is important and in your case it reports your cards right.

i am going to reboot this tomorrow so when i do i am going to go over the settings in cmos. presently it is set to auto on pci-e bus frequency. maybe i will fix it at 100mhz .. it could be doing God knows what in auto.

I don't think pci-e bus frequency has any noticeable effect in cuda speed and even if it did, it should affect both your cards, not only one.

also the 3 digit time workunits are still the 285 and 2 digit the tesla.  i wonder if it has something to do with how busy my desktops are? i have quite a lot going on 24/7 with 18 gkrellm server monitors running in one desktop, usually 4 or 5 browser windows in different desktops with maybe 28 or so tabs open, average 8 or 10 ssh konqueror tabs open into our servers, email, virtualbox running xp which also runs boinc, kopete, 8 or 9 postit notes in the various desktops, a few kedit windows open plus momentary things like adobe reader, smplayer or whatever.. im in totally new territory here. my experience in graphics cards is plug it in and make sure it works with a stable and peppy screen

however the 'busyness' of the desktops is not new and was basically the same when i had 10-13min workunits out of both cards.

This is a very busy desktop. Have you tried running a few workunits with absolutely nothing of the above running? I think your times will return to "normal".

Hello,
I try Crunch3rs CUDA seti application and search google and this forum too but result is not OK.
I tried to CUDA 2.1 and aplication setiathome-CUDA-6.08.x86_64-pc-linux-gnu. This compute OK but take 100% of CPU.
http://setiathome.berkeley.edu/result.php?resultid=1340190227

Now I test CUDA 2.3 with same apliaction and result is SIGSEGV: segmentation violation

I try add setiathome-CUDA-6.08.x86_64-pc-linux-gnu to /usr/local/bin but still not working.
CUDA 2.2 take segmentation violation too.
Thanks for any ideas
Libor

Cuda 2.1 libraries have a bug, that's why you have 100% CPU utilisation. You need 2.2 or later. Also please follow my post in http://lunatics.kwsn.net/linux/seti-mb-cuda-for-linux.msg19014.html#msg19014 very carefully.

I try this application too but here is another problem
ldd setiathome-6.08.CUDA_2.2_x86_64-pc-linux-gnu
./setiathome-6.08.CUDA_2.2_x86_64-pc-linux-gnu: /usr/lib64/libstdc++.so.6: version `GLIBCXX_3.4.9' not found (required by ./setiathome-6.08.CUDA_2.2_x86_64-pc-linux-gnu)
CentOS5 have no GLIBCXX_3.4.9 in updates now.
Libor

Centos, since it's more enterprise oriented, uses old versions of ...well everything. Have you had any success?

@riofl and lordvader about kernel versions
Have you compiled these kernels yourselves or have you got them from elsewhere? Maybe some performance/optimization options you left out? Do you have any nvidia related errors in your syslog when running cuda? Any other observations with these newer kernels?

Is it even possible to use CUDA with two different devices?  If so, what am I missing?

Of course it is. Any link to your host? Also try putting

<use_all_gpus>1</use_all_gpus>

in the options section of your cc_config.xml and tell boinc to read the config file.

[riofl]

i always compile my own kernels and have been since i first ran linux 8 yrs ago. in switching back and forth there was no difference in my cuda processing times. at first i thought there was but it was just different size workunits.

no, i have not tried just running boinc without a gui.. i will try that this coming weekend when i can spare some downtime from work and monitoring the servers. will let it run for 1 hr with no X running and then will go in and see if there are any differences.   

thing is, the usage of my desktops has not changed much at all during the past year so i had the same stuff open with the 13min workunits a few months ago. will be interesting to see if the 3 digit numbers move into 2 digit though on the tasks report.

thanks for the info about video settings. i was not sure what they did but when i did not see any changes, i figured "off" may be better..  i hate powermizer myself but i cannot find any options to turn it off and leave the card in high perf mode at all times. every time i spot check it its always in hi perf mode so maybe my temps are not high enough to trigger it (assuming temp is its onlly trigger) and if idle is a trigger, my desktop is never idle even when i go to bed, all the gkrellm monitors are advancing their graphs every second.

seems so strange with all the mb servers down, my cuda cards are both idling at around 46c. really odd since i am used to them being in the low or mid 60s all the time.

[sunu]

no, i have not tried just running boinc without a gui.. i will try that this coming weekend when i can spare some downtime from work and monitoring the servers. will let it run for 1 hr with no X running and then will go in and see if there are any differences.   

Leave X, just close all those apps you have running. Just the desktop with boinc in the background.

thing is, the usage of my desktops has not changed much at all during the past year so i had the same stuff open with the 13min workunits a few months ago. will be interesting to see if the 3 digit numbers move into 2 digit though on the tasks report.

The bigger multibeam workunits started about a month or two ago.

i hate powermizer myself but i cannot find any options to turn it off and leave the card in high perf mode at all times. every time i spot check it its always in hi perf mode so maybe my temps are not high enough to trigger it (assuming temp is its onlly trigger) and if idle is a trigger, my desktop is never idle even when i go to bed, all the gkrellm monitors are advancing their graphs every second.

Many people have tried many ways to turn off powermizer usually with no success.    Powermizer levels are triggered by GPU usage or very high (95+°C) temperatures.

seems so strange with all the mb servers down, my cuda cards are both idling at around 46c. really odd since i am used to them being in the low or mid 60s all the time.

I have some WUs cached for a few days more 

[riofl]

no, i have not tried just running boinc without a gui.. i will try that this coming weekend when i can spare some downtime from work and monitoring the servers. will let it run for 1 hr with no X running and then will go in and see if there are any differences.   

Leave X, just close all those apps you have running. Just the desktop with boinc in the background.

ok ill close down all my 'server' functions as well like my jabber server, bind, etc. so its just x and boinc running.

thing is, the usage of my desktops has not changed much at all during the past year so i had the same stuff open with the 13min workunits a few months ago. will be interesting to see if the 3 digit numbers move into 2 digit though on the tasks report.

The bigger multibeam workunits started about a month or two ago.

hehe thats about the time i started noticing issues. maybe they're not issues afer all.

i hate powermizer myself but i cannot find any options to turn it off and leave the card in high perf mode at all times. every time i spot check it its always in hi perf mode so maybe my temps are not high enough to trigger it (assuming temp is its onlly trigger) and if idle is a trigger, my desktop is never idle even when i go to bed, all the gkrellm monitors are advancing their graphs every second.

Many people have tried many ways to turn off powermizer usually with no success.    Powermizer levels are triggered by GPU usage or very high (95+°C) temperatures.

ok well i hardly do anything involving true graphics besides cuda running on that stuff and i have my hardware monitors set to shut the system down if the gpu gets to 80c.. once i adjusted the fans and air flow in the case they have never gone above 70c.

seems so strange with all the mb servers down, my cuda cards are both idling at around 46c. really odd since i am used to them being in the low or mid 60s all the time.

I have some WUs cached for a few days more 

lucky. im set for 10 days but ran out of cuda.. still have 400+ cpu units which are all vlar/vhar. been trying to get more since saturday.

i'll just wait out the stampede and give the servers a chance to settle and then go for it again.

[riofl]

ugh i have to learn to remember to add extra quote tags.. all my responses are in the quote area..

[sunu]

You can switch those VHARs to your graphics cards.

[riofl]

You can switch those VHARs to your graphics cards.

cool. i have more than enough of those

Number of CPU tasks:420
Number of GPU tasks:0
Number of VLAR tasks:103
Number of VHAR tasks:317

just changed the script around to give all vhar tasks to gpu so now cpu will only get vlar tasks plus spillover if any. looks like they are loving it too

Number of CPU tasks:103
Number of GPU tasks:317
Number of VLAR tasks:103
Number of VHAR tasks:317
Total tasks: 420


thanks!


interesting. ... on the vhar ones my gtx285 is finishing the work units 2 full minutes ahead of my tesla in every case... so far they finished 6 of them varying from  11 to 14  min each. will check out the tasks report to see if they say something different too, just for the info. dunno what it will prove if anything... the test over the weekend should tell whether im keeping the 285 a bit busy or not.... i would think the 285 would not even notice desktop activity...

[sunu]

I've never used that rebranding script and don't know how it works. Seeing the code I think changing

if($trueAR < 0.13 || $trueAR > 1.127)

to

if($trueAR < 1.127)

should do the trick. Backup first!

[riofl]

yep thats exactly what line i changed... simply commented out the original line and inserted the new line under it.

only i used the left side since the logic assignes that evaluation to cpu.

if($trueAR < 0.13){
     $tasks{$WUname}=603;

maybe to get a bit better distribution i need to change the number from 0.13 to maybe 0.5 to make it more fair for the cpus to get some non vlar work too?  think i will leave it like this and see what the distribution looks like over this week then maybe play with that number a bit keeping it below 0.60 to be safe..