You are not logged in.
I just started, gonna leave it on 24/7 unless it reduces my productivity
Offline
wow, i'm gonna need some 9800 gtx's in these machines to even be competitive lol. Didn't realize how much those things could crunch...... sorta sad that a graphics card can crunch more than my processors, 64 bit hyperthreaded 3.6 ghz processors at that. May go see what the local best buy has available so i can throw in a couple cards and improve my herd a bit.
Offline
Here is a thread that lists PPD vs PPD per Watt for various GPUs.
Here is a thread that lists PPD for the various GPUs.
Know any "gotta have the latest and greatest" type buddies. I have picked up some cheap 8800GT that friends were selling so they could buy GTXwhatever cards. Check Craigs list in your area or the local classified ads. You would be surprised what is out there at cheap prices with the recession going on. 8800GT, 9800GT, and 9800GTX all do well for me, don't have any GTXwhatevers yet.
Pudge
Offline
Here is a thread that lists PPD vs PPD per Watt for various GPUs.
Here is a thread that lists PPD for the various GPUs.
Know any "gotta have the latest and greatest" type buddies. I have picked up some cheap 8800GT that friends were selling so they could buy GTXwhatever cards. Check Craigs list in your area or the local classified ads. You would be surprised what is out there at cheap prices with the recession going on. 8800GT, 9800GT, and 9800GTX all do well for me, don't have any GTXwhatevers yet.
Pudge
Very interesting Pudge.
I am having lots of dependency problems with my system from adding and updating new camera software (just got a Canon 500D) and other kde stuff. Anyhow I am going to stop my gpu client until cuda, nvidia driver et al catch up...:(
Diesel1.
PS. When is your interview with ALM?
Registered GNU/Linux user #140607.
Offline
just hit 200,000 1million here i come
Desktop: E8400@4ghz - DFI Lanparty JR P45-T2RS - 4gb ddr2 800 - 30gb OCZ Vertex - Geforce 8800 GTS - 2*19" LCD
Server/Media Zotac GeForce 9300-ITX I-E - E5200 - 4gb Ram - 2* ecogreen F2 1.5tb - 1* wd green 500gb - PicoPSU 150xt - rtorrent - xbmc - ipazzport remote - 42" LCD
Offline
Nice 10 MILLION Pudge!
Congratulations, surely this means you get a 'STAR ARCHER' interview on the magazine now!
Diesel1.
Registered GNU/Linux user #140607.
Offline
I fired up one of my Arch/WINE/GPU computers, and did a pacman -Syu. Tried to start the GPU client and it said I had an Unstable machine. Same symptoms you guys are having.
Next, I removed nvidia-185.18.14-1 and nvidia-utils-185.18.14-1 and I removed lib32-nvidia-utils-185.18.14-1. I had to use pacman -Rd nvidia-utils to remove nvidia-utils because of existing dependencies. Using pacman -U I down graded the kernel to kernel26 2.6.29.4-1
Then used pacman -U to load nvidia-180.51-1 and nvidia-utils-180.51-1 and lib32-nvidia-utils-180.51-1. I had all these files in
/var/cache/pacman/pkg as I haven't cleared the pacman cache for a while.Rebooted, and the GPU client started up fine and is currently at 14% on the Work Unit.
Basically, my Arch 64 installation is totally up to date including the rpcbind thing, except for kernel26-2.6.29.4 and the 180.51 nvidia drivers. So, obviously there is a problem in one or both of them and down grading them will work around the problem for now.
HTH
Pudge
UPDATE:
I recently did a new install of ArchLinux 64 and set up the Windows 32 bit GPU client. I used the newest version of everything, WINE from the AUR, kernel, Nvidia drivers, and Cuda toolkit. The 2.6.30 series of kernel and the 185 series Nvidia drivers still do not work. But now we have to add something else to the list. You have to use Version 2.2 of the CUDA Toolkit. Version 2.3 will not work. In summary, to get the GPU client to work in Arch, you must use the following package versions:
kernel26 2.6.29.4-1 nvidia-180.51-1 nvidia-utils-180.51-1 lib32-nvidia-utils-180.51-1 CUDA Tookit Ver. 2.2
It doesn't seem to care what version of WINE you use.
Pudge
EDIT: For more information on this problem, see page 15 of this thread, starting around post 364.
Last edited by Pudge (2009-09-16 02:05:55)
Offline
The Archlinux Folding @ Home team now has eight Millionaires. Congratulations to kth5 for cranking out one million points.
Cheers.
Pudge
Offline
Nice MILLION kth5!
Diesel1.
Registered GNU/Linux user #140607.
Offline
Unfortunately I've been away for a long while, so I haven't been contributing anything. However, I have my computer set back up, and have been using kernel26-bfs from AUR for over a week now. It has been great when compiling things, it maxes out all four of my cores with -j4 without any hiccups anywhere else. So I figured, why not try and use it and only one SMP client this time? It just downloaded its first WU and is running at 100% on all four cores. It brings me great joy seeing my processor actually put to use, not to mention that my computer is still completely responsive! I'll report back and alert anybody of any problems I have with it. Has anybody else tried to use BFS with this?
Offline
Unfortunately I've been away for a long while, so I haven't been contributing anything. However, I have my computer set back up, and have been using kernel26-bfs from AUR for over a week now. It has been great when compiling things, it maxes out all four of my cores with -j4 without any hiccups anywhere else. So I figured, why not try and use it and only one SMP client this time? It just downloaded its first WU and is running at 100% on all four cores. It brings me great joy seeing my processor actually put to use, not to mention that my computer is still completely responsive! I'll report back and alert anybody of any problems I have with it. Has anybody else tried to use BFS with this?
what does the bfs kernel do different from stock?
i no it uses BFS scheduler but what does the scheduler do different?
think they should change the name lol
Last edited by markp1989 (2009-10-11 14:07:25)
Desktop: E8400@4ghz - DFI Lanparty JR P45-T2RS - 4gb ddr2 800 - 30gb OCZ Vertex - Geforce 8800 GTS - 2*19" LCD
Server/Media Zotac GeForce 9300-ITX I-E - E5200 - 4gb Ram - 2* ecogreen F2 1.5tb - 1* wd green 500gb - PicoPSU 150xt - rtorrent - xbmc - ipazzport remote - 42" LCD
Offline
I'm by no means an expert on kernel schedulers, and while I believe I have a relatively good grasp on what it does, you should see http://bbs.archlinux.org/viewtopic.php?id=79398 and http://ck.kolivas.org/patches/bfs/bfs-faq.txt so that I don't mis-explain something important. To me the specifics of what it does aren't the most important aspect, but rather that it works so well on my computer without the need to tune it. I've never looked into tuning the default scheduler (CFS?), but BFS definitely works great so far.
Offline
Well, I'm happy to report that my computer finished its first work unit in the middle of the night and sent it in without any problems. It shows up on Stanford's website as well. After that, in the last six hours or so, it has completed nearly half of another WU, which seems to be snappier but I don't really remember the exact values from before. Is anyone running a Q6600 on CFS that has any numbers to compare?
Also, Pudge, it is your post that suggest that people run more than just one client. Can you confirm that the only purpose for that was to have all cores on the CPU maxed out? If that is the only reason, I definitely suggest people try out BFS, it is simply amazing!
Offline
jac;
Without BFS running only one client on four cores will yield about 50% usage on all four cores. Running two instances of the client maxes out all four cores.
I doubt that I will try out BFS as all my four core Arch Linux computers are running the Windows GPU client through WINE besides the SMP client(s). The only way I am keeping them running now is with the 6.26.29 kernel and older Nvidia drivers. Throwing BFS into the mix probably wouldn't be wise at this time.
Glad BFS worked out for you, and I will definately keep it mind for the future.
Pudge
Offline
Brought my main rig online for the Arch f@h team a few hours ago. Running the smp client and a gpu client through wine.
I haven't tried BFS on my main rig due to the same reasons given by Pudge, but on the rigs I have tried it on it didn't seem to make much difference in folding speed. I stick with a 2.6.29 vanilla kernel with a minimal config and pretty much stay pegged at 100% minus 13% on one core for the GPU client. Only time I had trouble with not loading all cores was on 2665 WU's but they just don't scale with smp.
I'll try to bring another rig or two into the fray in a week or so.
Offline
I put my new computer on our F@H team a few days ago, with an overclocked GTX 275 (running in Windows, so I can play my video games) and two or three SMP clients running on a Core i7 920 in VMWare with Notfred's awesome F@H disk image. Up to 11k ppd now, definitely a great improvement over my old single-core Athlon 64 3700+.
If I get my HTPC up and running properly anytime soon, I might have it fold too; it's got an Athlon 64 3800+ in it and a Radeon 2400. Do Radeons work well with GPU folding in Linux?
...with a crowbar.
Offline
ATI GPU through wine is not working yet. There was a thread started at the foldingforum, some people were trying to get it going but they lacked core developer support and I haven't heard anything about since.
Offline
ATI GPU through wine is not working yet. There was a thread started at the foldingforum, some people were trying to get it going but they lacked core developer support and I haven't heard anything about since.
That's a shame... I think a single SMP client on my processor is only going to add a few hundred PPD.
...with a crowbar.
Offline
I'm back in. Using a FreeBSD server to contribute to the Arch team
Offline
The PlayStation3 have about 1 Million, you cant never beat this^^
Offline
That's a shame... I think a single SMP client on my processor is only going to add a few hundred PPD.
On a dual core CPU, a single SMP Client will average about 1,000 to 1,400 points a day. Also, no contribution is too small. Every Work Unit that gets processed is yet another step in the long journey towards a worthy goal.
The PlayStation3 have about 1 Million, you cant never beat this^^
It doesn't matter. Awarding points for WUs and forming teams, etc. is just playing on the competitive nature of humans. There is no denying that this competition keeps a lot of people interested in folding, and increases the number of WU folded. However, I don't think anyone should feel that their contribution is not worthy. The bottom line is, if you feel good about trying to give something back to society, then do so. I can't compete with a lot of personal contributions or coporate contributions during the Jerry Lewis MD telethon, but that doesn't stop me from putting a few dollars in a fireman's boot to help the cause.
P.S. I sure hope that this isn't taken as a rant, because it wasn't meant to be.
Pudge
Offline
JooBlixlarz wrote:That's a shame... I think a single SMP client on my processor is only going to add a few hundred PPD.
On a dual core CPU, a single SMP Client will average about 1,000 to 1,400 points a day. Also, no contribution is too small. Every Work Unit that gets processed is yet another step in the long journey towards a worthy goal.
Vamp898 wrote:The PlayStation3 have about 1 Million, you cant never beat this^^
It doesn't matter. Awarding points for WUs and forming teams, etc. is just playing on the competitive nature of humans. There is no denying that this competition keeps a lot of people interested in folding, and increases the number of WU folded. However, I don't think anyone should feel that their contribution is not worthy. The bottom line is, if you feel good about trying to give something back to society, then do so. I can't compete with a lot of personal contributions or coporate contributions during the Jerry Lewis MD telethon, but that doesn't stop me from putting a few dollars in a fireman's boot to help the cause.
P.S. I sure hope that this isn't taken as a rant, because it wasn't meant to be.
Pudge
You are absolutely right Pudge. Even though I'm not running the GPU client and getting a 'gazillion' points a day anymore, I am happy to contribute every WU I can.
Diesel1.
Registered GNU/Linux user #140607.
Offline
quad3datwork;
Welcome back to F@H. Glad to see you back.
Pudge
Offline
I've now finally got my server up and running, and it's now folding 24/7. Downside is that it's darn slow, 11% on 4 days, while my main computer, running about 20/7 are already finished with it's first since installation 4 days ago.
Offline
Thanks!
Added few Intel Q series, Xeons, and AMD Opterons. Fastest is Xeon X5570 (Dell R710 w/ 48GB RAM).
quad3datwork;
Welcome back to F@H. Glad to see you back.
Pudge
Last edited by quad3d@work (2009-10-26 14:46:44)
Offline