Advanced search

Message boards : News : WU: A2AR

Author Message
Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 35904 - Posted: 26 Mar 2014 | 15:13:32 UTC

A big hi to all the users! I'm currently undergoing my master thesis in the lab and I'll very proud to participate in this fantastic community for the first time by sending some simulations to gpugrid.

You will identify them with the ID A2AR, which stands for A2A adenosine receptor. It is a membrane protein from to a huge family called GPCRs that is key on neurodegenerative diseases such as Parkinson and therefore an important target of drug development. With these simulations we will try to prove the affinity of a couple of known ligands in silico.

TJ
Send message
Joined: 26 Jun 09
Posts: 815
Credit: 1,469,215,105
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35908 - Posted: 26 Mar 2014 | 16:16:02 UTC - in response to Message 35904.

Welcome Gerard, this research is also very important, so I am happy to crunch these. I have now one of them on my slow rig.
However my main reason for helping here is research and cures for cancer.
____________
Greetings from TJ

Snow Crash
Send message
Joined: 4 Apr 09
Posts: 450
Credit: 539,316,349
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35910 - Posted: 26 Mar 2014 | 16:36:33 UTC

Welcome to the forums :-) we're always happy to help process the work and really appreciate hearing about the work you are doing - crunchers always like to hear what the smart women and men in the lab are up to <thumbsup>
____________
Thanks - Steve

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 35915 - Posted: 26 Mar 2014 | 16:58:25 UTC

I forgot to note that due to the nature of these simulations, some small forces have to be added externally and, unfortunatenly, these have to be calculated using cpu instead of gpu. Therefore you may notice some amount of cpu usage that in my case never surpassed a 10%. I added a 20% credit bonus to encourage crunching. Hopefuly no errors will derive but if they do don't hesitate to post them!

KSUMatt
Avatar
Send message
Joined: 11 Jan 13
Posts: 214
Credit: 831,004,493
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwat
Message 35924 - Posted: 27 Mar 2014 | 2:30:09 UTC

Two completed successfully.

8150882 5468944 143481 26 Mar 2014 | 17:06:23 UTC 27 Mar 2014 | 0:08:07 UTC Completed and validated 18,753.96 6,986.43 135,000.00 Long runs (8-12 hours on fastest card) v8.15 (cuda55)

8150783 5468868 143481 26 Mar 2014 | 17:05:36 UTC 27 Mar 2014 | 0:35:01 UTC Completed and validated 18,747.59 6,989.17 135,000.00 Long runs (8-12 hours on fastest card) v8.15 (cuda55)
____________

biodoc
Send message
Joined: 26 Aug 08
Posts: 89
Credit: 656,130,328
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35925 - Posted: 27 Mar 2014 | 9:12:50 UTC

I've got 2 of these WUs running on 2 different machines. Both machines are running 64-bit linux mint 16. Once computer has a GTX680Ti and the other has a GTX570. The WU on the 680Ti shows less than 5% complete after nearly 10 hours. CPU usage is ~1% and the GPU temp is only 45C. Since it's linux I can't tell you GPU load but since the temp is low, the GPU load must also be very low. My NVIDIA driver version is 319.32.

The GTX570 on the other computer has the same issue: Less than 1% complete after 2 hours. Low CPU usage and low GPU temps.

Both computers have 1 core free for GPUGrid and usually use close to 100% of that core.

Here's a link to the computer with the GTX680Ti: http://www.gpugrid.net/results.php?hostid=154778


Any thoughts on what the problem is?

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 810,073,458
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 35932 - Posted: 27 Mar 2014 | 17:06:56 UTC

Happy to see another neuro project here. Great :)
____________
DSKAG Austria Research Team: http://www.research.dskag.at



Crunching for my deceased Dog who had "good" Braincancer..

KSUMatt
Avatar
Send message
Joined: 11 Jan 13
Posts: 214
Credit: 831,004,493
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwat
Message 35936 - Posted: 27 Mar 2014 | 18:15:11 UTC

I'm seeing this project actually has very low CPU usage compared to other GPUGrid tasks - less than a full CPU thread per GPU on 332.21 as you can see from the times I posted.
____________

Jacob Klein
Send message
Joined: 11 Oct 08
Posts: 1067
Credit: 1,145,477,339
RAC: 1,117,306
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35955 - Posted: 28 Mar 2014 | 11:02:27 UTC - in response to Message 35936.
Last modified: 28 Mar 2014 | 11:06:07 UTC

I'm seeing this project actually has very low CPU usage compared to other GPUGrid tasks - less than a full CPU thread per GPU on 332.21 as you can see from the times I posted.


We have already confirmed that, with the recent NVIDIA driver updates, we can expect the acemd application to not require a full CPU core anymore, even on a Kepler. According to the GPUGrid admins, the new drivers (including your version I believe) now honor the request from the GPUGrid application to run the app in a "low-CPU-utilization mode".

So, again, seeing the acemd app use only a third of a core... is now normal, even for Keplers.

Profile robertmiles
Send message
Joined: 16 Apr 09
Posts: 402
Credit: 169,015,396
RAC: 339,121
Level
Ile
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35998 - Posted: 30 Mar 2014 | 2:38:08 UTC - in response to Message 35904.

A big hi to all the users! I'm currently undergoing my master thesis in the lab and I'll very proud to participate in this fantastic community for the first time by sending some simulations to gpugrid.

You will identify them with the ID A2AR, which stands for A2A adenosine receptor. It is a membrane protein from to a huge family called GPCRs that is key on neurodegenerative diseases such as Parkinson and therefore an important target of drug development. With these simulations we will try to prove the affinity of a couple of known ligands in silico.


I've seen some articles elsewhere indicating that it may be significant whether certain brain proteins have an affinity for aluminum, copper or iron that is not already bound to some other protein, so are those among the known ligands you are testing? If not, do the GPUGRID software even allow testing them for affinity?

The human brain's natural defense against Alzheimer's is a small protein that also regulates the level of free copper ions inside the brain. Iron is known to bind to damaged tissues in the brain, and appears likely to be the cause of the brain damage in Parkinson's. Aluminum in known to bind to certain areas of the brain during a later stage of Alzheimer's - the same areas that are damaged in that stage.

FSU Scientist Links Iron Imbalance To Parkinson's Disease
http://www.sciencedaily.com/releases/2004/10/041013084413.htm

biodoc
Send message
Joined: 26 Aug 08
Posts: 89
Credit: 656,130,328
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36006 - Posted: 30 Mar 2014 | 20:57:57 UTC - in response to Message 35925.

I've got 2 of these WUs running on 2 different machines. Both machines are running 64-bit linux mint 16. Once computer has a GTX680Ti and the other has a GTX570. The WU on the 680Ti shows less than 5% complete after nearly 10 hours. CPU usage is ~1% and the GPU temp is only 45C. Since it's linux I can't tell you GPU load but since the temp is low, the GPU load must also be very low. My NVIDIA driver version is 319.32.

The GTX570 on the other computer has the same issue: Less than 1% complete after 2 hours. Low CPU usage and low GPU temps.

Both computers have 1 core free for GPUGrid and usually use close to 100% of that core.

Here's a link to the computer with the GTX680Ti: http://www.gpugrid.net/results.php?hostid=154778


Any thoughts on what the problem is?


If anyone's interested, upgrading to the latest Nvidia drivers solved my problem.

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 810,073,458
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 36585 - Posted: 23 Apr 2014 | 11:56:56 UTC
Last modified: 23 Apr 2014 | 12:01:49 UTC

Oh i see now these are the first ones witch are limiting on a pcie1 x4 slot on this project ^^ 570GTX on 90% while the other 570 in the x16 slot is at 99% . But i will soon install newer drivers (still the 301.xx ones) so i will see if it stays so. It it stays at 90% its not that bad, im happy i still can crunch good times with old hardware and soon an aditional 580gtx will follow :)
____________
DSKAG Austria Research Team: http://www.research.dskag.at



Crunching for my deceased Dog who had "good" Braincancer..

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 810,073,458
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 36604 - Posted: 24 Apr 2014 | 7:32:50 UTC
Last modified: 24 Apr 2014 | 7:33:48 UTC

Oh i wrote something wrong, it istn bottlenecked by x4 slots because i see the 90% now on the x16 slot too, seems to be an cpu power issue or perhaps still the driver ^^ its only a q6600. Only to correct my previous post..
____________
DSKAG Austria Research Team: http://www.research.dskag.at



Crunching for my deceased Dog who had "good" Braincancer..

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 36624 - Posted: 24 Apr 2014 | 18:30:17 UTC
Last modified: 24 Apr 2014 | 18:42:36 UTC

First results from this research line. Seems that we got lucky! check it out: https://www.youtube.com/watch?v=e84fbaJbw1g

Peter_M
Send message
Joined: 25 Feb 14
Posts: 15
Credit: 23,570,837
RAC: 0
Level
Pro
Scientific publications
watwatwatwat
Message 36647 - Posted: 25 Apr 2014 | 17:37:33 UTC - in response to Message 36006.


If anyone's interested, upgrading to the latest Nvidia drivers solved my problem.


I'm very interested, thank you biodoc, what driver version exactly are you running ? I have the same issue, around 5% both for CPU and GPU usage on a GTX 750Ti with Drivers 334.21 on Ubuntu 12.04.4 and the WU makes almost no progress. I already installed the patched boinc client as suggested in this thread but no difference.

biodoc
Send message
Joined: 26 Aug 08
Posts: 89
Credit: 656,130,328
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36649 - Posted: 25 Apr 2014 | 18:01:19 UTC - in response to Message 36647.


If anyone's interested, upgrading to the latest Nvidia drivers solved my problem.


I'm very interested, thank you biodoc, what driver version exactly are you running ? I have the same issue, around 5% both for CPU and GPU usage on a GTX 750Ti with Drivers 334.21 on Ubuntu 12.04.4 and the WU makes almost no progress. I already installed the patched boinc client as suggested in this thread but no difference.


I'm running the latest beta driver, 337.12. I hope that works for you.

Vagelis Giannadakis
Send message
Joined: 5 May 13
Posts: 187
Credit: 349,254,454
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwat
Message 36669 - Posted: 26 Apr 2014 | 13:05:13 UTC - in response to Message 36624.

I won't pretend that I understand much from the video you linked Gerard, but I'm happy that we're making progress! ("we" as in mainly you with our help)

Thanks for the update and keep up the good work!
____________

Vagelis Giannadakis
Send message
Joined: 5 May 13
Posts: 187
Credit: 349,254,454
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwat
Message 36825 - Posted: 13 May 2014 | 8:17:42 UTC

I just aborted A2ART4E_adapt1_01x7-GERARD_A2ART4E_adapt1-7-17-RND5409_0. It had reached 4.33% after 7h:17m and completion estimation had risen to more than 168h! The GPU (750 Ti) was almost idle and CPU utilization was 0-1% (for this task). I judged it to be basically stuck, so I aborted it.

Unfortunately, its output does not contain much...
____________

Vagelis Giannadakis
Send message
Joined: 5 May 13
Posts: 187
Credit: 349,254,454
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwat
Message 36858 - Posted: 17 May 2014 | 16:17:02 UTC

My GTX 750 TI refused to process nicely yet another GERARD_A2ART4E_adapt2. The symptom is the same every time: progresses 2-3% and then seems to proceed at an extremely slow rate, sky-rocketing completion estimations to ~168 hours. The GPU gets close to idle (solely judging by its temperature since I'm on Linux).
____________

Jim1348
Send message
Joined: 28 Jul 12
Posts: 460
Credit: 1,130,761,180
RAC: 22,822
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwat
Message 36859 - Posted: 17 May 2014 | 16:43:25 UTC - in response to Message 36858.

For whatever it is worth to know, my two GTX 750 Ti's have processed all eight A2ARs that they have received with no problems or resets, under WinXP (335.28 driver). Maybe there are some driver problems with other OS's?
http://www.gpugrid.net/results.php?hostid=173205&offset=0&show_names=1&state=0&appid=

But I don't restart this machine, and it has nothing else running (not even other BOINC projects), so that might help.

Profile Carlesa25
Avatar
Send message
Joined: 13 Nov 10
Posts: 324
Credit: 72,394,453
RAC: 0
Level
Thr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36861 - Posted: 17 May 2014 | 16:58:54 UTC

Hi: I just finished a A2AR task smoothly on my Ubuntu 14.04 Nvidia 337.19. Greetings.

Peter_M
Send message
Joined: 25 Feb 14
Posts: 15
Credit: 23,570,837
RAC: 0
Level
Pro
Scientific publications
watwatwatwat
Message 36865 - Posted: 19 May 2014 | 6:15:10 UTC - in response to Message 36858.

My GTX 750 TI refused to process nicely yet another GERARD_A2ART4E_adapt2. The symptom is the same every time: progresses 2-3% and then seems to proceed at an extremely slow rate, sky-rocketing completion estimations to ~168 hours. The GPU gets close to idle (solely judging by its temperature since I'm on Linux).


Same here, GX750Ti and Ubuntu 12.04, thus I usually cancel these jobs, if I get them. Others (Barna, Nathan etc) run nice, so no problem for me. Just to confirm you are not alone :)

ETQuestor
Send message
Joined: 11 Jul 09
Posts: 26
Credit: 777,242,386
RAC: 567,075
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37080 - Posted: 19 Jun 2014 | 20:58:59 UTC - in response to Message 36858.

I am having the same issue - A2AR WUs running very slowly - in fact, too slowly to finish before the reporting deadline. I also have a GTX 750Ti, running on Fedora 20 x86_64, with NVIDIA 337.25 x86_64 driver.

HA-SOFT, s.r.o.
Send message
Joined: 3 Oct 11
Posts: 100
Credit: 4,855,582,826
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37083 - Posted: 20 Jun 2014 | 8:25:17 UTC - in response to Message 37080.
Last modified: 20 Jun 2014 | 8:26:47 UTC

Same here. On 750ti and linux and win machines. Progres goes normal to ~0.700 then return to zero and cycles and after some time very slowly increases ~1% in 5 hours.

Vagelis Giannadakis
Send message
Joined: 5 May 13
Posts: 187
Credit: 349,254,454
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwat
Message 37084 - Posted: 20 Jun 2014 | 8:54:22 UTC

Yesterday my 750Ti (on Ubuntu 12.04) completed A2AR_standAlonex37-GERARD_A2AR_standAlone-14-17-RND0620_0 successfully. It's the first such WU I see completed on it, although I have aborted quite a few since my first ~10 A2ARs stalled. Maybe something has changed on some new batch(es)?
____________

HA-SOFT, s.r.o.
Send message
Joined: 3 Oct 11
Posts: 100
Credit: 4,855,582,826
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37086 - Posted: 20 Jun 2014 | 10:50:03 UTC - in response to Message 37084.

I think it depends on wu. Not all wu's have this problem.

Michael P. Gainor
Send message
Joined: 23 Jan 16
Posts: 15
Credit: 69,388,188
RAC: 0
Level
Thr
Scientific publications
wat
Message 42740 - Posted: 2 Feb 2016 | 0:43:29 UTC

Hello, Gerard I saw my GPU was also working on some of your CXCl_12 does this research having anything to do with your A2AR or something else? In either case I would be love if you could elaborate on this work that is under CXCl_12.

Post to thread

Message boards : News : WU: A2AR