Advanced search

Message boards : Server and website : No jobs for the GPU

Author Message
ElleSolomina
Avatar
Send message
Joined: 22 Mar 14
Posts: 43
Credit: 527,549,678
RAC: 14,812
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 46433 - Posted: 4 Feb 2017 | 8:30:08 UTC
Last modified: 4 Feb 2017 | 8:30:55 UTC

there are no available jobs. Generally there are no jobs. There is no short or long for the GPU. It's in my computer a problem with getting jobs or something with the project?

http://www.gpugrid.net/show_host_detail.php?hostid=170640

http://www.gpugrid.net/show_host_detail.php?hostid=170784
____________

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 6,169
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46434 - Posted: 4 Feb 2017 | 9:59:56 UTC - in response to Message 46433.

It's the project. Take a look at the server status page (top right corner).
The current simulations reached their final steps (workunits generated from finished tasks here at GPUGrid), so there's only very scarce work available for a while until the staff will release new simulations.

Bedrich Hajek
Send message
Joined: 28 Mar 09
Posts: 467
Credit: 8,195,021,966
RAC: 10,422,078
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46435 - Posted: 4 Feb 2017 | 12:09:17 UTC

About a month ago, we had a bunch of WUs (SDOERR_CASP22S) posted, then canceled, but never re-released. I wonder what happened to them?

I still have 5 ghost units, on one of my computers from that time.




Stefan
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 5 Mar 13
Posts: 348
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 46444 - Posted: 6 Feb 2017 | 8:42:46 UTC - in response to Message 46435.

The CASP ones are the equivalent of the current ADRIA_FOLD ones. We will look this week at the results from the short queue and decide what to send out so we might send a bigger batch again.

3de64piB5uZAS6SUNt1GFDU9d...
Avatar
Send message
Joined: 20 Apr 15
Posts: 285
Credit: 1,102,216,607
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwat
Message 46445 - Posted: 6 Feb 2017 | 9:51:35 UTC

May I repeat my previous proposal ... if we cannot increase the number of studies for using GPUGRID to capacity, why don't we aim at larger and more sophisticated molecules?

I seem to recal a F@H posting where the scientists stated that truly thrilling investigations stand and fall with crunching power and it was hardly possible to derive something valuable before the 2010s because of lacking speed. The hardware was simply not fast enough to handle larger proteins.

Well, here we have a supercomputer now stuffed with all kinds of Kepler, Maxwell and Pascal GPUs running idle... and craving for big data.
____________
I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 46446 - Posted: 6 Feb 2017 | 11:49:51 UTC

Perhaps they could shorten the length of the WUs, I say this even as a fast card owner. The work just isn't being done efficiently and with more WUs more people could have a chance to contribute in parallel rather than waiting for a gt640 to finish a BNBS2.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 46447 - Posted: 6 Feb 2017 | 17:59:26 UTC

The new ADRIA_FAAH_WT_ Long WUs only run about 75% on my 970 on windows 10. The Bus interface load is less than BNBS2 and Pablo's adaptive goal. Not sure if you guys can improve this. Haven't gotten one of these new ones on my windows XP 64bit 980ti yet so I can't say if WDDM has much difference.

Erich56
Send message
Joined: 1 Jan 15
Posts: 1090
Credit: 6,603,906,926
RAC: 18,783,925
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwat
Message 46449 - Posted: 6 Feb 2017 | 19:36:12 UTC - in response to Message 46447.
Last modified: 6 Feb 2017 | 19:36:34 UTC

Haven't gotten one of these new ones on my windows XP 64bit 980ti yet so I can't say if WDDM has much difference.

Supposedly it makes some difference, but not to the full extent (i.e. the difference between 75% and full load).

ElleSolomina
Avatar
Send message
Joined: 22 Mar 14
Posts: 43
Credit: 527,549,678
RAC: 14,812
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 46452 - Posted: 7 Feb 2017 | 5:17:17 UTC - in response to Message 46434.

It's the project. Take a look at the server status page (top right corner).
The current simulations reached their final steps (workunits generated from finished tasks here at GPUGrid), so there's only very scarce work available for a while until the staff will release new simulations.


Thank you! I will wait for the emergence of new jobs. It is a pity that the job ended now, because now in my region the coldest time. In the summer, for the northern hemisphere it is unlikely that I will take on the GPU.
____________

EMYArg
Send message
Joined: 5 Apr 12
Posts: 32
Credit: 381,502,763
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46478 - Posted: 8 Feb 2017 | 20:35:44 UTC

I see two topics that talk about the same...
Can project managers tell us when there will be new research and new tasks?
Thank you.

Dj
Send message
Joined: 15 Jan 17
Posts: 1
Credit: 92,979,142
RAC: 0
Level
Thr
Scientific publications
watwatwat
Message 46482 - Posted: 9 Feb 2017 | 14:34:47 UTC

I certainly would appreciate being able to get more than one long run and one short run every 3-5 days. I bought a new GPU specifically so that this project would no longer take nearly the deadline time to process one long run WU, now that I can do a long run in about 7 hours, having no GPU work for the same amount of time my old system used to crunch one WU isn't something I appreciate. :(

Einstein is getting a lot of Priority-Zero work from me with the new GPU though. I just wish there were more GPU projects for bio/medical sciences... or more consistently available WU's here!

kain
Send message
Joined: 3 Sep 14
Posts: 152
Credit: 641,182,245
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 46483 - Posted: 9 Feb 2017 | 16:08:00 UTC

There is folding@home :)

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 46494 - Posted: 11 Feb 2017 | 13:41:37 UTC

6th day in a row none of my computers have gotten a single WU, despite manually updating when the new WUs were released. The event log never said too many updates, it just spit out the normal 0 new tasks sent.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 46495 - Posted: 11 Feb 2017 | 13:43:46 UTC
Last modified: 11 Feb 2017 | 13:47:06 UTC

It might have something to do with the fact that all of them are doing Einstein both CPU and GPU just trying to keep my house warm and perhaps because of this they don't ask for GPUGrid work as often, is there anyway I can change this?

3de64piB5uZAS6SUNt1GFDU9d...
Avatar
Send message
Joined: 20 Apr 15
Posts: 285
Credit: 1,102,216,607
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwat
Message 46496 - Posted: 11 Feb 2017 | 14:55:12 UTC - in response to Message 46483.

There is folding@home :)

+1
____________
I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

jjch
Send message
Joined: 10 Nov 13
Posts: 98
Credit: 15,288,150,388
RAC: 1,732,962
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46498 - Posted: 11 Feb 2017 | 21:40:46 UTC
Last modified: 11 Feb 2017 | 21:41:13 UTC

PappaLitto - Set your GPUGRID Resource share to the maximum of 99,999. That will give this project the highest possible priority in BOINC.

If you have other GPU projects you want to run while waiting for GPUGRID work set the Resource shares lower. I have Milkyway set to 500, Einstein set to 100 and Seti set to 0. The same is also true for CPU projects but they are two separate groups of work so the relative priority is split between CPU and GPU groups.

Also set your work queue low so it doesn't download lots of work from other projects that would stack up and excessively delay starting a GPUGRID work unit when one comes in. Mine is set at .05 days work and .05 additional. That gives a little buffer but not more than 2.4 hours total.

Folding is another option however it is completely separate from BOINC. It is possible to run them both on the same computer and have the folding application run on the GPU only and use BOINC for the CPU only projects. I have a major investment in BOINC so I decided to stick with that.

kain
Send message
Joined: 3 Sep 14
Posts: 152
Credit: 641,182,245
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 46499 - Posted: 11 Feb 2017 | 22:12:32 UTC - in response to Message 46498.

I have a major investment in BOINC so I decided to stick with that.


What exactly?

jjch
Send message
Joined: 10 Nov 13
Posts: 98
Credit: 15,288,150,388
RAC: 1,732,962
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46500 - Posted: 12 Feb 2017 | 6:54:37 UTC

I was referring to my time investment with setting up systems running BOINC, configuring and optimizing them. I also am using BoincStats to monitor and control these systems. Another consideration is I am participating in the Gridcoin group and I didn't see a way to use folding@home with the same setup.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 46505 - Posted: 14 Feb 2017 | 13:29:12 UTC

I think it's extremely silly that folding@home doesn't also put themselves on BOINC. The entire reason I use BOINC is the ability to have backup projects and Gridcoin. Folding could significantly increase their FLOPS if they supported BOINC.

Erich56
Send message
Joined: 1 Jan 15
Posts: 1090
Credit: 6,603,906,926
RAC: 18,783,925
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwat
Message 46506 - Posted: 14 Feb 2017 | 17:56:28 UTC - in response to Message 46505.

I think it's extremely silly that folding@home doesn't also put themselves on BOINC.

I remember that some time ago I have read somehwere (most likely on the folding@home HP) that despite all efforts taken, for some technical reasons it did not work out, as much as the FH people would have liked FH to be running under BOINC.

Really too bad.

Jim1348
Send message
Joined: 28 Jul 12
Posts: 819
Credit: 1,591,285,971
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46507 - Posted: 14 Feb 2017 | 18:53:17 UTC - in response to Message 46506.

It is because Folding needs lower latency. That is, the output of one completed work unit is needed to start work on the next in the series. So they can't take the long buffer times in BOINC. That is why the Folding client starts downloading a new work unit when the present one reaches 99% complete.

I don't quite know why BOINC couldn't do that with a little tweaking though, but I don't think anyone has looked into it for several years. BOINC actually has some advantages; it is able to report more information to the server on the driver you are using, for example, and a lot of other information about your hardware. Also, BOINC detects your GPU cards properly, and allows you some degree of individual control. That has been a never-ending headache with Folding, though they are attempting to fix it now with their new client, currently in Beta. We will see.

Betting Slip
Send message
Joined: 5 Jan 09
Posts: 670
Credit: 2,498,095,550
RAC: 0
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46510 - Posted: 15 Feb 2017 | 12:16:44 UTC - in response to Message 46507.

It is because Folding needs lower latency. That is, the output of one completed work unit is needed to start work on the next in the series. So they can't take the long buffer times in BOINC. That is why the Folding client starts downloading a new work unit when the present one reaches 99% complete.


This project needs lower latency as well in order to provide a more contiguos flow of work.
I proposed a similar strategy in this post but while I was hopeful in its implementation that hope is now evaporating.

https://gpugrid.net/forum_thread.php?id=4501

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 6,169
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46511 - Posted: 15 Feb 2017 | 13:43:57 UTC - in response to Message 46510.

It is because Folding needs lower latency. That is, the output of one completed work unit is needed to start work on the next in the series. So they can't take the long buffer times in BOINC. That is why the Folding client starts downloading a new work unit when the present one reaches 99% complete.

This project needs lower latency as well in order to provide a more contiguos flow of work.
I proposed a similar strategy in this post but while I was hopeful in its implementation that hope is now evaporating.
https://gpugrid.net/forum_thread.php?id=4501

It's up to the BOINC developers, or this project should be outside of BOINC just like folding@home. It would be sufficient to have different queue length in BOINC for different projects to achieve this behavior. This queue length even could be set to a fixed amount by the project itself.
But it's possible even now to achieve this behavior, if you set a low queue length (0.01 days = 14 minutes 24 seconds), and make the BOINC manager to estimate the remaining time on the progress of the workunit. (As this project does not set correct values for the <rsc_fpop_est> field in a workunit's properties.)

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 46512 - Posted: 15 Feb 2017 | 15:22:42 UTC

Because of the lack of work, I have put all of my processors on folding@home. I let BOINC run with network activity on always in the background and manually shut off folding when it eventually downloads a GPUGrid WU.

John C MacAlister
Send message
Joined: 17 Feb 13
Posts: 181
Credit: 144,871,276
RAC: 0
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 46518 - Posted: 17 Feb 2017 | 15:09:13 UTC - in response to Message 46512.

Because of the lack of work, I have put all of my processors on folding@home. I let BOINC run with network activity on always in the background and manually shut off folding when it eventually downloads a GPUGrid WU.


Working on FAH, childhood cancer, cancer mapping, rosetta, asteroids, zika, TB.

I will return to GPUGrid as soon as there is work.


____________
John

kain
Send message
Joined: 3 Sep 14
Posts: 152
Credit: 641,182,245
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 46527 - Posted: 18 Feb 2017 | 14:27:49 UTC

When can we expect normal supply?

3de64piB5uZAS6SUNt1GFDU9d...
Avatar
Send message
Joined: 20 Apr 15
Posts: 285
Credit: 1,102,216,607
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwat
Message 46538 - Posted: 22 Feb 2017 | 18:37:51 UTC



;-))
____________
I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

JDragonB
Send message
Joined: 5 Dec 16
Posts: 3
Credit: 2,313,210
RAC: 0
Level
Ala
Scientific publications
watwat
Message 46563 - Posted: 26 Feb 2017 | 9:45:46 UTC

Stefan
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 5 Mar 13
Posts: 348
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 46615 - Posted: 9 Mar 2017 | 14:43:13 UTC

How about now? ;)

3de64piB5uZAS6SUNt1GFDU9d...
Avatar
Send message
Joined: 20 Apr 15
Posts: 285
Credit: 1,102,216,607
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwat
Message 46616 - Posted: 9 Mar 2017 | 15:40:50 UTC
Last modified: 9 Mar 2017 | 15:46:43 UTC

Thanks! MUCH better, I love it. Keep it up please!
Besides, does the server now respond faster ... or is that just my imagination?
____________
I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

lohphat
Send message
Joined: 21 Jan 10
Posts: 44
Credit: 704,692,359
RAC: 1,198,087
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46774 - Posted: 25 Mar 2017 | 4:07:42 UTC

25mar2017:

Tasks ready to send 2,847
Tasks in progress 4,100

Yet when I force a project update, I get no tasks.

I've joined other projects which use my GPU instead because GPUGRID is unreliable at issuing work units.

Jacob Klein
Send message
Joined: 11 Oct 08
Posts: 1127
Credit: 1,901,927,545
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 46788 - Posted: 26 Mar 2017 | 13:42:21 UTC - in response to Message 46774.
Last modified: 26 Mar 2017 | 13:42:38 UTC

25mar2017:

Tasks ready to send 2,847
Tasks in progress 4,100

Yet when I force a project update, I get no tasks.

I've joined other projects which use my GPU instead because GPUGRID is unreliable at issuing work units.


- If your GPUs already have enough work to meet your setting for "Store at least X days of work", then it won't get more.
- If GPUGrid is not the next project in line to receive work, according to your Recent Estimated Credit (REC), then it won't get any.
- If you'd like to figure out the real problem, turn on Options > Event Log Options... > Work Fetch Debug, and then read through the Event Log to see WHY you weren't given work, and try to step through that logic by looking at the log.
- If you need help, post one iteration of that work fetch debug logic where you thought it was suspicious, and someone here might help you further.

Post to thread

Message boards : Server and website : No jobs for the GPU

//