Advanced search

Message boards : News : Maxwell GTX750Ti

Author Message
Profile MJH
Project administrator
Project developer
Project scientist
Send message
Joined: 12 Nov 07
Posts: 696
Credit: 27,266,655
RAC: 0
Level
Val
Scientific publications
watwat
Message 35518 - Posted: 5 Mar 2014 | 11:54:13 UTC

Hi Guys,

Just a line to say we've got our first GTX750Ti card up and running. The performance out-of-the-box matches a GTX480. We'll have a public app for GPUGRID just as soon as CUDA 6.0 is publicly released, which will probably happen at GTC at the end of the month.

Matt

Profile Mumak
Avatar
Send message
Joined: 7 Dec 12
Posts: 92
Credit: 225,897,225
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwat
Message 35519 - Posted: 5 Mar 2014 | 12:19:20 UTC - in response to Message 35518.
Last modified: 5 Mar 2014 | 12:20:07 UTC

Thanks for the info.
Can you please post a few results (time) how it performs here ? I assume "up and running" means that it works with some sort of your beta CUDA 6 app...

Profile GDF
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 14 Mar 07
Posts: 1895
Credit: 629,356
RAC: 0
Level
Gly
Scientific publications
watwatwatwatwat
Message 35522 - Posted: 5 Mar 2014 | 14:13:33 UTC - in response to Message 35519.
Last modified: 5 Mar 2014 | 14:14:13 UTC

a cuda6 application will be available asap. Minimum driver required 334.21.

gdf

Profile MJH
Project administrator
Project developer
Project scientist
Send message
Joined: 12 Nov 07
Posts: 696
Credit: 27,266,655
RAC: 0
Level
Val
Scientific publications
watwat
Message 35526 - Posted: 5 Mar 2014 | 16:16:01 UTC - in response to Message 35519.


Can you please post a few results (time) how it performs here ?


Sorry, the most I can say is that it's comparable to a GTX480 - someone crunching with one of those will have to tell you what RAC that translates to.


I assume "up and running" means that it works with some sort of your beta CUDA 6 app...


That's right - we have the Release Candidate of CUDA 6 here in the lab. It's not a public release yet, so we're not at liberty to distribute the runtime libraries that our application needs.

Matt

Snow Crash
Send message
Joined: 4 Apr 09
Posts: 450
Credit: 539,316,349
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35529 - Posted: 6 Mar 2014 | 0:12:35 UTC
Last modified: 6 Mar 2014 | 0:14:34 UTC

Running SANTI WUs a GTX480 will get about ~230K-~250K PPD on Win7. With a decent OC you can get it up to ~275K. Other types of WUs will get a bit lower to a bit higher (NOELIAs can get you up to about 300K PPD).
____________
Thanks - Steve

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 29,512
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35757 - Posted: 20 Mar 2014 | 9:50:57 UTC

Got a first cuda6 long run wu
http://www.gpugrid.net/workunit.php?wuid=5301509

Runtime was 46030 sec; setup includes running of igpu wu's (2 Einstein)
As tested earlier, the benefit for the GTX750ti of running no igpu wu's is in the range of 10% the conclusion is: the GTX750ti can crunch 2 long runs (36x2478-NOELIA_FXArep-0-2-RND0793) in one day.

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,835,229,774
RAC: 291,018
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35758 - Posted: 20 Mar 2014 | 9:57:58 UTC - in response to Message 35757.

Thanks Alexander,
That could provide a daily RAC of ~253,000 (while crunching Einstein iGPU WU's) or around 278,000 if not.
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Profile FourOh
Send message
Joined: 20 Feb 14
Posts: 3
Credit: 100,000,070
RAC: 0
Level
Cys
Scientific publications
watwatwatwatwatwat
Message 35869 - Posted: 24 Mar 2014 | 19:46:48 UTC

My 750 Ti is getting Cuda55 long work units and they all error out... the short units are Cuda60 and are successful. A few days ago I got some long run Cuda60 units which were also sucessful - but I'm not getting those anymore.

Is there anything I can do to force Cuda 6 long run units?

http://www.gpugrid.net/results.php?hostid=170512

Peter_M
Send message
Joined: 25 Feb 14
Posts: 15
Credit: 23,570,837
RAC: 0
Level
Pro
Scientific publications
watwatwatwat
Message 36077 - Posted: 3 Apr 2014 | 16:33:31 UTC

Hello,
I'm crunching for a while but I'm new to GPUGrid, I run a GTX 750Ti on Ubuntu 12.04.4 using 334.21 NVIDIA drivers in various GPU projects. Is it still valid that there are no or only rarely WU's for this card ? I have GPU grid added since about half a day and all I get is the message in the Messages log that there are no WU's available.

Profile MJH
Project administrator
Project developer
Project scientist
Send message
Joined: 12 Nov 07
Posts: 696
Credit: 27,266,655
RAC: 0
Level
Val
Scientific publications
watwat
Message 36078 - Posted: 3 Apr 2014 | 16:37:48 UTC - in response to Message 36077.

Peter_M: You should be getting WUs like any other client...

Peter_M
Send message
Joined: 25 Feb 14
Posts: 15
Credit: 23,570,837
RAC: 0
Level
Pro
Scientific publications
watwatwatwat
Message 36079 - Posted: 3 Apr 2014 | 16:47:11 UTC

Thanks,
and sorry for being in the "News" section, my fault. All I get is this

Do 03 Apr 2014 18:44:07 CEST | GPUGRID | Sending scheduler request: To fetch work.
Do 03 Apr 2014 18:44:07 CEST | GPUGRID | Requesting new tasks for NVIDIA
Do 03 Apr 2014 18:44:08 CEST | GPUGRID | Scheduler request completed: got 0 new tasks
Do 03 Apr 2014 18:44:08 CEST | GPUGRID | No tasks sent
Do 03 Apr 2014 18:44:08 CEST | GPUGRID | No tasks are available for Short runs (2-3 hours on fastest card)
Do 03 Apr 2014 18:44:08 CEST | GPUGRID | No tasks are available for ACEMD beta version
Do 03 Apr 2014 18:44:08 CEST | GPUGRID | No tasks are available for Long runs (8-12 hours on fastest card)

any advice or pointer where to look into ?

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2671
Credit: 751,855,924
RAC: 450,091
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36081 - Posted: 3 Apr 2014 | 21:26:02 UTC

Has the linux app already been updated for Maxwell? Otherwise the recent scheduler changes may prevent sending the wrong app. Or something simply went wrong with that scheduler fix.

MrS
____________
Scanning for our furry friends since Jan 2002

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 791
Credit: 1,425,102,570
RAC: 1,358,780
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36082 - Posted: 3 Apr 2014 | 21:46:06 UTC - in response to Message 36081.

Has the linux app already been updated for Maxwell? Otherwise the recent scheduler changes may prevent sending the wrong app. Or something simply went wrong with that scheduler fix.

MrS

I don't see a cuda60 for Linux on the applications page.

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,835,229,774
RAC: 291,018
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36089 - Posted: 4 Apr 2014 | 22:23:09 UTC - in response to Message 36082.
Last modified: 4 Apr 2014 | 22:27:13 UTC

There is only a CUDA6 app for Linux in the Beta queue,

    Linux running on an AMD x86_64 or Intel EM64T CPU 8.11 (cuda60) 15 Mar 2014 | 0:07:21 UTC


However, going by the Server status page there are no Beta tasks,

    ACEMD beta version 0 0 0.00 (0.00 - 0.00) 0


...So there are no CUDA 6 tasks for Linux!
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Peter_M
Send message
Joined: 25 Feb 14
Posts: 15
Credit: 23,570,837
RAC: 0
Level
Pro
Scientific publications
watwatwatwat
Message 36094 - Posted: 5 Apr 2014 | 11:00:00 UTC - in response to Message 36089.

However, going by the Server status page there are no Beta tasks,
    ACEMD beta version 0 0 0.00 (0.00 - 0.00) 0


...So there are no CUDA 6 tasks for Linux!



Ok, thank you, so not necessarely something wrong with my setup. Thank you everyone for your help! Have a wonderful weekend,
Peter

Profile MJH
Project administrator
Project developer
Project scientist
Send message
Joined: 12 Nov 07
Posts: 696
Credit: 27,266,655
RAC: 0
Level
Val
Scientific publications
watwat
Message 36097 - Posted: 5 Apr 2014 | 12:11:50 UTC - in response to Message 36094.
Last modified: 5 Apr 2014 | 13:57:07 UTC

Correct - I need to build a CUDA 6 app for Linux.

Matt

Profile MJH
Project administrator
Project developer
Project scientist
Send message
Joined: 12 Nov 07
Posts: 696
Credit: 27,266,655
RAC: 0
Level
Val
Scientific publications
watwat
Message 36100 - Posted: 5 Apr 2014 | 14:23:13 UTC - in response to Message 36097.

Linux app CUDA 6.0 is now available for cc5.0 / GTX750 etc.

Matt

Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 211
Credit: 12,256,290,796
RAC: 8,451,959
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36102 - Posted: 5 Apr 2014 | 17:54:53 UTC
Last modified: 5 Apr 2014 | 18:21:57 UTC

On my Linux machines, cuda 6.0 wu's are being sent to every card, regardless of driver compatibility & consequently all failing. Oh dear.

Driver 331.49 is the minimum req for cuda 6.0 & with that driver it has not failed (yet) :-)

Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 211
Credit: 12,256,290,796
RAC: 8,451,959
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36103 - Posted: 5 Apr 2014 | 19:29:35 UTC

Nuts!!! Even the short queue is infested with cuda6 wu's. No option but to stop receiving work for now on most of my machines :-(

Profile MJH
Project administrator
Project developer
Project scientist
Send message
Joined: 12 Nov 07
Posts: 696
Credit: 27,266,655
RAC: 0
Level
Val
Scientific publications
watwat
Message 36104 - Posted: 5 Apr 2014 | 19:58:13 UTC - in response to Message 36103.

Huh. That's unexpected. What driver version do your machines have?

Matt

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 791
Credit: 1,425,102,570
RAC: 1,358,780
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36105 - Posted: 5 Apr 2014 | 20:16:49 UTC - in response to Message 36104.

Huh. That's unexpected. What driver version do your machines have?

Matt

One of the problems may be that BOINC for Linux never seems to have been able to extract the version numbers from CUDA drivers and report it to the servers.

Profile MJH
Project administrator
Project developer
Project scientist
Send message
Joined: 12 Nov 07
Posts: 696
Credit: 27,266,655
RAC: 0
Level
Val
Scientific publications
watwat
Message 36106 - Posted: 5 Apr 2014 | 20:20:48 UTC - in response to Message 36105.


One of the problems may be that BOINC for Linux never seems to have been able to extract the version numbers from CUDA drivers and report it to the servers.


Well, that might account for it, I'll have to see what values the scheduler is working with.

In any case, Stoneageman: try version 821, to rule out a possible app build problem.

Matt

Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 211
Credit: 12,256,290,796
RAC: 8,451,959
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36107 - Posted: 5 Apr 2014 | 20:21:54 UTC - in response to Message 36104.
Last modified: 5 Apr 2014 | 20:23:28 UTC

Huh. That's unexpected. What driver version do your machines have?

Matt

304.88 for my 500 & 600 series cards & 331.20 for the 780's. I'm now trying 331.49 on a 780 with cuda6 & it looks to be at least 5% slower than 331.20.

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,835,229,774
RAC: 291,018
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36108 - Posted: 5 Apr 2014 | 21:15:47 UTC - in response to Message 36107.
Last modified: 5 Apr 2014 | 21:16:06 UTC

With the recent drivers, performance can be capped due to a lack of GPU utilization; I'm presently running a GERRARD Long run task on CUDA 6 (albeit on a GTX770 and W7) and the clocks are 1045MHz. Power is at 40%. It's normally OC'ed to ~1250MHz. Even with only 3 CPU threads in use for CPU projects and the GPU Utilization rising to 80% the GPU remains in that non boost clock rate (it drops but doesn't rise). It's a known issue, but it's hard to spot - might not happen for days if you only receive some WU types. So far I have only seen this on the first GPU, and rarely. This driver 'feature' might also be applicable to Linux.

I'm using 331.49 on Linux with a GTX670 and a GTX660Ti. I get the odd failure on that system, but so far I've not seen any issues with the few tasks that ran on CUDA6. Recently completed a Gianni WU using v8.20 (cuda6.0), presently running a NATHAN_KID using 8.20 (cuda6.0).
Going by NVidia X server the GTX670 GPU clock varies from 1071 to 1110MHz (with Prefer Maximum Performance selected).

I have not received any cuda6 WU's on my other Linux system with an earlier driver.
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 211
Credit: 12,256,290,796
RAC: 8,451,959
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36109 - Posted: 5 Apr 2014 | 21:31:18 UTC
Last modified: 5 Apr 2014 | 21:32:05 UTC

I don't believe it! Just installed 331.49 on my 4x780 machine & now it's downloading 8x cuda5.5 tasks :-(

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,835,229,774
RAC: 291,018
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36111 - Posted: 5 Apr 2014 | 22:05:18 UTC - in response to Message 36109.
Last modified: 5 Apr 2014 | 22:05:53 UTC

Was going to suggest using the Beta queue, but at least they won't all fail overnight!
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 211
Credit: 12,256,290,796
RAC: 8,451,959
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36113 - Posted: 5 Apr 2014 | 22:52:06 UTC

Looks like Matt has updated the cuda 6 app from 8.20 to 8.21

Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 211
Credit: 12,256,290,796
RAC: 8,451,959
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36118 - Posted: 6 Apr 2014 | 12:19:52 UTC - in response to Message 36108.

With the recent drivers, performance can be capped due to a lack of GPU utilization

Bring back swan_sync :-)
I've gone back to 331.20 & now just getting the occasional cuda6 task.

Jacob Klein
Send message
Joined: 11 Oct 08
Posts: 1068
Credit: 1,148,170,114
RAC: 1,026,871
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36119 - Posted: 6 Apr 2014 | 12:37:59 UTC - in response to Message 36118.
Last modified: 6 Apr 2014 | 12:38:47 UTC

With the recent drivers, performance can be capped due to a lack of GPU utilization

Bring back swan_sync :-)
I've gone back to 331.20 & now just getting the occasional cuda6 task.


I don't think swan_sync has anything to do with maintaining max boost during situations where GPU utilization is low enough that it downclocks.

If you are looking at maintaining max boost, please have a look at my thread here, which describes a method to do it on Windows.
http://www.gpugrid.net/forum_thread.php?id=3647
It might be adapted to be usable for linux, somehow, if you can find a way to force the P2 power state.

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 791
Credit: 1,425,102,570
RAC: 1,358,780
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36135 - Posted: 7 Apr 2014 | 10:41:12 UTC - in response to Message 36113.

Looks like Matt has updated the cuda 6 app from 8.20 to 8.21

pvh in number crunching is reporting that v8.21 is failing all tasks on GTX 550 Ti and GTX 660 Ti with driver 331.20 and 64-bit Linux (openSUSE 13.1).

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,835,229,774
RAC: 291,018
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36160 - Posted: 8 Apr 2014 | 14:11:19 UTC - in response to Message 36135.

8.21 also failed on my Ubuntu 13.10 system (331.49 drivers and either or both of a GTX670 and a GTX660Ti). Failures were quick:

I435-NATHAN_KIDc22_full3-4-10-RND0271_0 5588452 159186 6 Apr 2014 | 18:33:13 UTC 6 Apr 2014 | 18:36:13 UTC Error while computing 79.69 19.83 --- Long runs (8-12 hours on fastest card) v8.21 (cuda60)

I1R7-NATHAN_KIDc22_glu3-4-10-RND0934_0 5588270 159186 6 Apr 2014 | 17:24:38 UTC 6 Apr 2014 | 18:33:13 UTC Error while computing 11.17 2.92 --- Long runs (8-12 hours on fastest card) v8.21 (cuda60)
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Post to thread

Message boards : News : Maxwell GTX750Ti