Posts by Pharrg

1) Questions and Answers : Windows : Can Rosetta and GPU version of Folding run at same time? (Message 60034)
Posted 9 Mar 2009 by Pharrg
Post:
Hi. I am running the GPU version of Folding@Home on my nVidia GTX260 graphics cards using CUDA. There is almost no use of my CPU's with that. Since they do not use BOINC, I was wondering if anyone had tried to let Rosetta@Home run via BOINC on the CPU's at the same time. That way, I can have both parts of my machine working and benefiting both projects. Anyone know if it would work? or will they interfere with each other?
2) Message boards : Number crunching : Problems with Minirosetta v1.54 (Message 59632)
Posted 17 Feb 2009 by Pharrg
Post:
So... I completed a bunch more tasks successfully, then got a 2nd task where it said the output file was missing. Anyone else getting these?

2/17/2009 6:20:35 AM rosetta@home Computation for task ss-neg-1i17__7365_5964_0 finished
2/17/2009 6:20:35 AM rosetta@home Output file ss-neg-1i17__7365_5964_0_0 for task ss-neg-1i17__7365_5964_0 absent

I noticed that both tasks that gave the 'absent output file' message had a name the started witht the same first part:

ss-neg-1i17__7365_

perhaps a bug in that one?
3) Message boards : Number crunching : Problems with Minirosetta v1.54 (Message 59626)
Posted 17 Feb 2009 by Pharrg
Post:
Ok, after a number of successful completions, I did see one that looks like it failed. Message as follows:

2/16/2009 7:49:12 PM rosetta@home Computation for task ss-neg-1i17__7365_4677_1 finished
2/16/2009 7:49:12 PM rosetta@home Output file ss-neg-1i17__7365_4677_1_0 for task ss-neg-1i17__7365_4677_1 absent


Don't know the cause of that one...
4) Message boards : Number crunching : Problems with Minirosetta v1.54 (Message 59625)
Posted 17 Feb 2009 by Pharrg
Post:
I started running Rosetta this morning on a 64bit Vista machine and all seems to be working well. It's been working well on other projects too. Here is what I'm running:

Core i7 920 CPU
Asus P6T6 WS Revolution motherboard
6Gb DDR3 Triple Channel RAM
Vista Home Premium SP1 64bit

64bit BOINC 6.6.7

As I said, no problems yet and a number of WU's have completed already.

5) Message boards : Number crunching : Minirosetta v1.54 Compute Errors (Message 59624)
Posted 17 Feb 2009 by Pharrg
Post:
I started running Rosetta this morning on a 64bit Vista machine and all seems to be working well. It's been working well on other projects too. Here is what I'm running:

Core i7 920 CPU
Asus P6T6 WS Revolution motherboard
6Gb DDR3 Triple Channel RAM
Vista Home Premium SP1 64bit

64bit BOINC 6.6.7

As I said, no problems yet and a number of WU's have completed already.

6) Message boards : Number crunching : A SINGLE Graphics Card = 1 teraFLOPS (Message 59623)
Posted 17 Feb 2009 by Pharrg
Post:
Oh, I forgot, in my limited testing with CUDA processing, I've traced pretty much all errors I've seen, including 'nvlddmkm' driver errors, driver recovery errors, BSOD's, and other crashes when trying to run CUDA to temperatures. This is especially a problem on the high end GTX 200 series of cards. Even with the cards being two slots thick, they just can't squeeze a big enough heatsink and fan into it. Just look at how large CPU coolers have gotten in comparison for much less computing power. I installed a Cooler Master V8 cooler on my Core i7 and that thing is a monster. But, I must admit, it works extremely well. In fact, it keeps my CPU cooler under a load than the stock cooler kept it at idle.

I found that once I maxed out my fan speeds and cooling on the video cards, the errors stopped. Many extreme gamers have run into the same issues on these boards as well. Lots of people try to blame the software or GPU's themselves, when really it's a simple case of overheating. Even games like Crysis with all settings maxed are nothing compared to what CUDA is capable of doing to a GPU when efficient code is used. The type of algorithms and processes will make a difference too, so you'll see different results from different projects and apps. It will drive your card hard. Like I said, once I figured this out and worked to keep my cards cool, I haven't had a crash since.

7) Message boards : Number crunching : A SINGLE Graphics Card = 1 teraFLOPS (Message 59622)
Posted 17 Feb 2009 by Pharrg
Post:
Wow! Must be nice. Of course, I don't have a $15,000 budget. However, I currently have a pair of GTX 260's, and the speed gain I've seen on some CUDA enable projects have been so impressive, I'm thinking of replacing them with 3 GTX295's. My motherboard will support true 3-way SLI with all 3 slots at full x16. Of course, when running CUDA, you have to disable SLI or only 1 card gets used, but that's a simple mouse click to do.

As for speeds.... 3 nVidia GTX 295's... Each board has 480 cores in parallel and almost 2Gb of dedicated DDR3 RAM. Each is capable of 1.788 Teraflops of processing power. Altogether, by running 3, this will give my machine 1440 cores and a whopping 5.364 Teraflops of power! That's faster than many early Cray supercomputers still in operation were! And, it will leave your CPU nearly idle for doing other tasks!

Do the math to see what his machine is going to be capable of... now that's extreme! Lots of great science can be done with these leaps of processing power. If AMD doesn't go bankrupt soon, perhaps the ATI Firestream technology will take off too, or perhaps OpenCL.


8) Message boards : Number crunching : Intel i7 CPU (Message 59621)
Posted 17 Feb 2009 by Pharrg
Post:
Hi,

If anyone is still wanting to look at Core i7 920 performance, I've just attached my recently built system to Rosetta today. There isn't much done yet since I just re-attached a few hours ago, but we'll see how it compares. Note that I have limited my CPU time to 90% because I don't like my CPU temperature to get over about 58c. I'm also using my pc while it's crunching, so that's probably impacting it some too. I do not overclock either. This machine wasn't cheap to build, so I don't want to toast it anytime soon! I'm also running a 64bit version of Vista and 64bit BOINC 6.6.7, so wonder if that will matter much. Anyway, here's a link to my machine if someone wants to monitor, though this page seems to show much less credit than BOINC manager does. Perhaps it is just delayed in updating.

http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=1008858

I'll let it run Rosetta for a day perhaps, but will likely then switch to one of the other projects so my CUDA cores can resume being used.
9) Message boards : Number crunching : Computing Preferences Page does not match (Message 59618)
Posted 17 Feb 2009 by Pharrg
Post:
Thanks. The reason I noticed this is because I had indeed updated my preferences on SaH. However, this morning when I re-attached to RaH, I noticed that it only retrieved a single task. When I looked at the messages tab, I saw a message saying number of usable cpu's changed from 8 to 1 due to preferences. So, I went to the Rosetta preferences page and noticed that it did in fact say 1 cpu even though I'd been using 8 on all the other projects since being on this new machine. When I looked closer at the page, I noticed many other differences. Hence this thread. I thought there might be another problem as well since it still wouldn't get any additional tasks even after I did an update, but I just got home from work and noticed it does now have 8 tasks running concurrently (Core i7 - 4 cores, 2 threads each). All seems well now.

I think the reason many of us like to have things consistant between projects, especially with preferences, is because now that some projects are CUDA enabled, we can have our CUDA devices running one project, and another running on the CPU. But that means we need to be able to utilize the GPU related options to do that. As it is now, when I'm running Rosetta, I now have 432 GPU CUDA enabled cores sitting idle while only 4 CPU cores are running. (I have two nVidia GTX 260 video cards.)

Anyway, thanks for the info...

10) Message boards : Number crunching : Computing Preferences Page does not match (Message 59607)
Posted 16 Feb 2009 by Pharrg
Post:
Your computing preferences page does not match the current page being used by other projects. There are a number of fields and options that are not there that should be. Since this page sets preferences for ALL projects, it needs to be updated. Look at the Rosetta computing preferences page, and compare to the SETI one, for example, here

http://setiathome.berkeley.edu/prefs.php?subset=global

It causes problems. For example, my system uses 8 cores, but because of the differences, when I used the Rosetta project today for the first time since building my machine, it changed the number of CPU's to 1. It didn't bring in some changes I'd made on other project sites to computing preferences, nor to the settings I make here get used in other projects, even though this page should be global. I noticed other issues too.






©2021 University of Washington
https://www.bakerlab.org