Advanced search

Message boards : News : New app on acemdbeta with Maxwell support

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 35581 - Posted: 10 Mar 2014 | 10:58:32 UTC

There's a new application on acemdbeta, version 811 for CUDA 60 which supports the Geforce 750Ti. You'll receive this app if your driver version is 334.21 or greater.

Please report problems here.

Matt

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35586 - Posted: 10 Mar 2014 | 17:08:41 UTC - in response to Message 35581.
Last modified: 10 Mar 2014 | 17:13:34 UTC

334.x or above is required to support the 750Ti and 750, but I would suggest trying 335.23 if anyone has a GTX750Ti or GTX750 because the driver brings "Enhanced GPU clock offset options for GeForce GTX 750 Ti and GeForce GTX 750". Let's hope Boost isn't still broken.
____________
FAQ's

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

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35589 - Posted: 10 Mar 2014 | 18:10:32 UTC - in response to Message 35581.

There's a new application on acemdbeta, version 811 for CUDA 60 which supports the Geforce 750Ti. You'll receive this app if your driver version is 334.21 or greater.

Please report problems here.

Matt



Hi,
I have driver 334.89 installed, enabled GPUGRID again, but got a cuda5.5 wu, which failed with errorcode -9 (0xfffffffffffffff7) Unknown error number.

Beta-apps enabled.
Any advice?

Alexander

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35590 - Posted: 10 Mar 2014 | 18:51:23 UTC - in response to Message 35589.

I've updated now to latest driver 335.23, got again a cuda 5.5 wu, which failed.

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 810,073,458
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 35591 - Posted: 10 Mar 2014 | 19:02:34 UTC

Your running short queue....


____________
DSKAG Austria Research Team: http://www.research.dskag.at



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

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35592 - Posted: 10 Mar 2014 | 19:12:11 UTC - in response to Message 35590.
Last modified: 10 Mar 2014 | 19:17:26 UTC

A CUDA6.0 Windows Beta app is now listed,



Alexander, You ran 3 short tasks,
Short runs (2-3 hours on fastest card) v8.15 (cuda55)

Possibly because you have selected to run other tasks.

To only get Beta's (when available),

Select,
Run test applications Yes
ACEMD beta: Yes

Deselect,
ACEMD short runs (2-3 hours on fastest card): No
[i]ACEMD long runs (8-12 hours on fastest GPU):
No
If no work for selected applications is available, accept work from other applications? No
Should not matter about the CPU app.

Looks like there are no Betas now,
ACEMD beta version 0 0 0.00 (0.00 - 0.00) 0
http://www.gpugrid.net/server_status.php
____________
FAQ's

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

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35593 - Posted: 10 Mar 2014 | 19:19:42 UTC - in response to Message 35591.

Your running short queue....



Right, but the original post did not say anything about short or long runs.
Anyway, I've changed my setup, enabled the long runs, enabled gpugrid again and got a ... cuda 55 app. A long run cuda 55.

Afterwards I changed my setup to allow beta apps only (because the original message says acemdbeta). Message from server: no tasks available.

Alexander

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35595 - Posted: 10 Mar 2014 | 19:38:55 UTC - in response to Message 35592.


To only get Beta's (when available),

Select,
Run test applications Yes
ACEMD beta: Yes

Deselect,
ACEMD short runs (2-3 hours on fastest card): No
[i]ACEMD long runs (8-12 hours on fastest GPU):
No
If no work for selected applications is available, accept work from other applications? No
Should not matter about the CPU app.

Looks like there are no Betas now,
ACEMD beta version 0 0 0.00 (0.00 - 0.00) 0
http://www.gpugrid.net/server_status.php


Thank you all for the help, my setup looks like you posted. Will wait to get a cuda 60 app.


Alexander

Profile FourOh
Send message
Joined: 20 Feb 14
Posts: 3
Credit: 100,000,070
RAC: 0
Level
Cys
Scientific publications
watwatwatwatwatwat
Message 35597 - Posted: 10 Mar 2014 | 21:05:10 UTC - in response to Message 35581.

What do you recommend in a mixed system (GTX 580 and GTX 750 Ti)?

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35601 - Posted: 10 Mar 2014 | 21:42:23 UTC - in response to Message 35597.
Last modified: 10 Mar 2014 | 21:42:57 UTC

What do you recommend in a mixed system (GTX 580 and GTX 750 Ti)?

A driver that supports your GTX750Ti, so probably the latest WHQL driver (335.23).
____________
FAQ's

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

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35603 - Posted: 10 Mar 2014 | 21:49:38 UTC - in response to Message 35601.

It would be useful if we could select to run Beta tasks and then either Long or Short tasks if Beta tasks are not available. At present you get both Long and Short if you select to run Other work if Beta's are not available.

- Just saying!
____________
FAQ's

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

Bedrich Hajek
Send message
Joined: 28 Mar 09
Posts: 340
Credit: 3,819,818,009
RAC: 929,462
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35604 - Posted: 11 Mar 2014 | 1:44:01 UTC

Ok, bring them on!

Profile Dingo
Avatar
Send message
Joined: 1 Nov 07
Posts: 11
Credit: 11,710,095
RAC: 9,029
Level
Pro
Scientific publications
watwatwatwatwatwatwatwat
Message 35645 - Posted: 14 Mar 2014 | 19:43:16 UTC
Last modified: 14 Mar 2014 | 20:10:49 UTC

I too have a GTX 750 TI and before I saw this post tried to run GPUGRID and the tasks all failed. I set my preferences to receive Beta and also the acemdbeta task but there are no tasks available.

I just updated to Driver version 335.23

Is there going to more of these tasks???
Cheers
____________

Proud Founder and member of



Have a look at my WebCam

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 35646 - Posted: 14 Mar 2014 | 21:32:46 UTC - in response to Message 35645.

I've put more WUs on for acemdbeta.

Matt

Bedrich Hajek
Send message
Joined: 28 Mar 09
Posts: 340
Credit: 3,819,818,009
RAC: 929,462
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35653 - Posted: 15 Mar 2014 | 4:29:56 UTC - in response to Message 35646.

I got a few of these units, so far they are running okay. They take about 2 hours to finish on my computers. Though, we get only 1500 credits for the effort. I don't see any noticeable difference in time to finish, between to cuda 6.0 v8.11 and cuda 5.5 v8.15, so far.

7928147 5280412 14 Mar 2014 | 21:09:36 UTC 14 Mar 2014 | 23:42:44 UTC Completed and validated 6,941.99 2,322.48 1,500.00 ACEMD beta version v8.11

7928277 5280531 15 Mar 2014 | 0:23:23 UTC 15 Mar 2014 | 3:37:59 UTC Completed and validated 6,968.59 2,386.53 1,500.00 ACEMD beta version v8.15 (cuda55)

http://www.gpugrid.net/results.php?hostid=127986&offset=0&show_names=0&state=0&appid=21

Bedrich Hajek
Send message
Joined: 28 Mar 09
Posts: 340
Credit: 3,819,818,009
RAC: 929,462
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35654 - Posted: 15 Mar 2014 | 6:02:30 UTC - in response to Message 35646.
Last modified: 15 Mar 2014 | 6:16:14 UTC

The output files for these units are very large (~45Mb).

So far, I have gotten only one cuda 6.0 WU on host 127986, the rest are either cuda 5.5 or 4.2. Why is that?

Why am I getting these units on a computer with driver: 327.23?

It was mentioned that driver: 335.23 was needed.



http://www.gpugrid.net/results.php?hostid=30790&offset=0&show_names=0&state=0&appid=21

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35659 - Posted: 15 Mar 2014 | 9:51:02 UTC - in response to Message 35654.
Last modified: 15 Mar 2014 | 9:59:04 UTC

On W7 GTX670 I've only seen 2 tasks run on the CUDA6 app,

7928183 5280447 14 Mar 2014 | 23:16:07 UTC 15 Mar 2014 | 1:44:46 UTC Completed and validated 6,236.75 1,973.52 1,500.00 ACEMD beta version v8.11 (cuda60)

7928159 5280424 14 Mar 2014 | 21:20:24 UTC 15 Mar 2014 | 0:00:19 UTC Completed and validated 6,204.24 1,951.20 1,500.00 ACEMD beta version v8.11 (cuda60)

Other than those 2 WU's I've only received cuda5.5 and cuda4.2 work for both Linux and Windows systems.

At present there isn't a Windows beta cuda6 application.
While there is one for Linux I haven't received any tasks (not that I have a Maxwell).
____________
FAQ's

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

Jim1348
Send message
Joined: 28 Jul 12
Posts: 460
Credit: 1,130,761,180
RAC: 18,722
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwat
Message 35660 - Posted: 15 Mar 2014 | 10:51:49 UTC

I am running WinXP on two different PCs, with three GTX 660s (335.xx drivers). One CUDA 4.2 and four CUDA 5.5 work units have completed successfully. However, the only CUDA 6.0 work unit failed right at the start (0 seconds):

GPUGRID 8.11 ACEMD beta version (cuda60) 76-MJHARVEY_TEST1999-0-10-RND4103_0 00:00:00 (00:00:00) 3/14/2014 11:45:02 PM 3/14/2014 11:48:03 PM 0.732C + 1NV 0.00 Reported: Computation error (313,) GPUGrid-PC

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 788
Credit: 1,422,060,845
RAC: 1,410,932
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35661 - Posted: 15 Mar 2014 | 12:44:02 UTC - in response to Message 35659.

At present there isn't a Windows beta cuda6 application.
While there is one for Linux I haven't received any tasks (not that I have a Maxwell).

Which is a little strange, because there was one last night - I processed task 7929440 using it. A couple of instabilities in the middle, but it completed OK.

Wdethomas
Send message
Joined: 6 Feb 10
Posts: 38
Credit: 274,204,838
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwat
Message 35669 - Posted: 15 Mar 2014 | 23:34:01 UTC - in response to Message 35661.
Last modified: 15 Mar 2014 | 23:37:35 UTC

Most of my downloads are the beta runs. Maybe only one long run. Just hope that my RAC does not go down because of these betas. I have under GPUGRID preference no beta.

UPDATE: I just changed to "no test work units", let's see.

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35675 - Posted: 16 Mar 2014 | 10:52:07 UTC - in response to Message 35669.
Last modified: 16 Mar 2014 | 10:56:37 UTC

Richard, when GPU's reach ~79 to 80°C some sort of throttling kicks in and the boost usually drops. Whatever is happening on the GPU at this time it causes tasks to become unstable. Fortunately the app can now recover the WU's, but the individuals solution is to increase GPU fan speeds, reducing the temperatures, so GPU's don't reach 79C. It's better than starting and stopping tasks and as it increases GPU longevity, decreases heat and power usage its the generic thing to do when GPU crunching.

Wdethomas, the short WU's and the Beta's yield less credits/time than the long WU's. If you want good credits just crunch the long WU's.

There is an 8.15 (cuda60) app in the shot queue for Windows, and an 8.11 (cuda60) app in the Beta queue for Linux. I would suggest that the CUDA 6 apps (also 4.2 and 4.5) work for non Maxwell GPU's (Kepler cards and below). I don't know if there is any point in people who don't have Maxwell's running these WU's?
____________
FAQ's

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

Wdethomas
Send message
Joined: 6 Feb 10
Posts: 38
Credit: 274,204,838
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwat
Message 35676 - Posted: 16 Mar 2014 | 11:15:07 UTC - in response to Message 35675.
Last modified: 16 Mar 2014 | 11:27:14 UTC

The Long Run 8.15 are generating only 54,300 points. I was getting 115,000 before. it seems that the driver 334 is affecting now the scoring.

It seems that the Titans are the ones generating these numbers . The Tesla are doing 115,000. Will keep an eye on this. Having the 4 cards on the same motherboard generates some drivers problems for me. Will be getting another motherboard to install the Teslas alone.

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35677 - Posted: 16 Mar 2014 | 11:52:43 UTC - in response to Message 35676.

The recent Long tasks that you ran were shorter than previous tasks:

e4s8_e1s4f56-GIANNI_trypben3-0-1-RND7147_1 5285887 16 Mar 2014 | 0:06:56 UTC 16 Mar 2014 | 4:12:39 UTC Completed and validated 8,179.43 3,027.14 54,300.00 Long runs (8-12 hours on fastest card) v8.15 (cuda55)

867x-SANTI_MAR420cap310-26-32-RND0600_1 5285097 16 Mar 2014 | 0:49:34 UTC 16 Mar 2014 | 8:58:10 UTC Completed and validated 24,500.15 5,792.68 115,650.00 Long runs (8-12 hours on fastest card) v8.15 (cuda55)

They are a different type of WU (from a different researcher).
Gianni's tasks give more credit per day, 86400/8179 * 54300 = 573,605 credits/day.
The Santi_Mar WU's gave, 86400/24500 * 115650 = 407,843 credits/day
____________
FAQ's

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

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 788
Credit: 1,422,060,845
RAC: 1,410,932
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35680 - Posted: 16 Mar 2014 | 12:10:42 UTC - in response to Message 35675.

Richard, when GPU's reach ~79 to 80°C some sort of throttling kicks in and the boost usually drops. Whatever is happening on the GPU at this time it causes tasks to become unstable. Fortunately the app can now recover the WU's, but the individuals solution is to increase GPU fan speeds, reducing the temperatures, so GPU's don't reach 79C. It's better than starting and stopping tasks and as it increases GPU longevity, decreases heat and power usage its the generic thing to do when GPU crunching.

In my case, the solution is to preferentially crunch GPUGrid tasks on the (lower, cooler) Device 1. Device 0 crunches other projects.

I was doing that manually, and the unexpected Beta cuda60 arrived late in the evening and ran overnight on Device 0.

I've now set up a project exclusion for device 0 in cc_config.xml, so that shouldn't happen again. I also agree that there isn't much point in running the Beta tasks using a non-Maxwell card and last November's applications, so I've disabled Beta apps for the time being. Happy to have verified that the cuda60 app does run on pre-Maxwell hardware, and I'll watch out for re-deployment.

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35684 - Posted: 16 Mar 2014 | 12:34:25 UTC

Most of the ACEMD beta version v8.15 (cuda55) I got failed immediate after starting.
http://www.gpugrid.net/results.php?userid=25200

I disabled these beta versions for now.

Alexander

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35685 - Posted: 16 Mar 2014 | 12:47:28 UTC - in response to Message 35684.

Alexander, there was a CUDA6 Beta app (in the Beta queue) for Windows, but it's not there now.
At present there is a CUDA6 app in the short queue for Windows. Perhaps you want to see if you get any CUDA6 work there? I'm not sure if there is any, and any CUDA55 or CUDA42 tasks will fail on a Maxwell, but I guess it's worth a shot.
____________
FAQ's

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

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35686 - Posted: 16 Mar 2014 | 14:12:09 UTC - in response to Message 35685.

Alexander, there was a CUDA6 Beta app (in the Beta queue) for Windows, but it's not there now.
At present there is a CUDA6 app in the short queue for Windows. Perhaps you want to see if you get any CUDA6 work there? I'm not sure if there is any, and any CUDA55 or CUDA42 tasks will fail on a Maxwell, but I guess it's worth a shot.


Ah, yes and no.
I can remember a post that said with driver xxx I will get the cuda 6 wu's. My mistake, I thought, I will get ONLY cuda6 wu's.
Anyway, I changed my setup and enabled the short runs and the beta's. I had the long runs enabled before.
Let's see what happens.

Alexander

PS: while writing these lines the download of my first cuda6 wu started!

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35690 - Posted: 16 Mar 2014 | 18:40:21 UTC - in response to Message 35686.

Many thanks Alexander:

818x-SANTI_MAR419cap310-29-84-RND6520_0 5288118 16 Mar 2014 | 14:01:45 UTC 16 Mar 2014 | 18:26:07 UTC Completed and validated 14,621.79 4,475.01 18,300.00 Short runs (2-3 hours on fastest card) v8.15 (cuda60)

Now we know it works, and it makes sense for the app to be in the short queue!
____________
FAQ's

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

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35693 - Posted: 16 Mar 2014 | 19:52:00 UTC - in response to Message 35690.

Many thanks Alexander:

818x-SANTI_MAR419cap310-29-84-RND6520_0 5288118 16 Mar 2014 | 14:01:45 UTC 16 Mar 2014 | 18:26:07 UTC Completed and validated 14,621.79 4,475.01 18,300.00 Short runs (2-3 hours on fastest card) v8.15 (cuda60)

Now we know it works, and it makes sense for the app to be in the short queue!


Glad I could help !
But the question remains; how can I make a setup to get the cuda6 wu's only? This should be a function of the server (as I understand it).

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35694 - Posted: 16 Mar 2014 | 20:11:55 UTC

Another question - but first a statement.
There are discussions over @ Einstein about gpu load and what it means. For all of the volunteers here, @ Einstein it's very common to run two or more wu's together on one gpu. For my GTX 750ti two wu's are quite goot, resulting in 95% gpu load and < 60% TDP power consumption, according to GPU-Z 0.7.7

The cuda-version that is used @ Einstein is cuda32.

Running one of the beta wu's here I see 78% gpu load, 300MB dedicated mem, 39MB dynamic mem, 22% mem controller load and up to 75% TDP power load.

There are two questions now:
- with a lower gpu usage - how can the power load be higher?
- to get a higher gpu load ( I am a fan of >95% ) - how can I run 2 wu's together? CPU cores are available, this is not a limit. Is there a chance to get a choice via the settings menue like Einstein offers ( the GPU utilization factor) ?

Alexander

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35697 - Posted: 16 Mar 2014 | 21:25:51 UTC - in response to Message 35694.
Last modified: 16 Mar 2014 | 21:28:09 UTC

Alexander, a couple of questions for you:
Where you running Einstein on the iGPU (i7-3770) when running the GPUGRID WU?
- edit - Looks like you were; (Binary Radio Pulsar Search (Arecibo) v1.34 (opencl-intel_gpu))

How many CPU cores where you using to crunch CPU projects when you completed the CUDA6 WU?
If you have GPUZ, open it and select the Sensors Tab, what is the Memory Controller Load when running a GPUGrid WU?

I agree that it's better to run 2 GPUGrid WU's at a time on Einstein, but as you point out it's a CUDA 3.2 app (due to bugs) and at GPUGrid more recent CUDA versions are used (4.2, 5.5 and now 6.0).

The complexity of the program determines actual power draw and GPU usage, as does the amount of memory and the number of shaders you use... It's also the case that CPU apps use different amounts of power. I once found that a Q6600's power usage could vary by >30W depending on what Boinc apps its running (running 4 CPU apps).

When the power usage is higher you are doing more 'Work' (Applied Mechanics).

Running more that 1 WU has been exhaustively discussed. Basically, you would need a high end 3GB card and very poor performing tasks to make it worthwhile here. On my Windows system I'm getting 85 to 92% GPU usage for my GPUGrid tasks depending on the WU type. Both my GPU's are overclocked by ~10%, almost eliminating the WDDM issues.
____________
FAQ's

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

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35700 - Posted: 16 Mar 2014 | 22:28:44 UTC - in response to Message 35697.

Alexander, a couple of questions for you:
Where you running Einstein on the iGPU (i7-3770) when running the GPUGRID WU?
- edit - Looks like you were; (Binary Radio Pulsar Search (Arecibo) v1.34 (opencl-intel_gpu))

How many CPU cores where you using to crunch CPU projects when you completed the CUDA6 WU?
If you have GPUZ, open it and select the Sensors Tab, what is the Memory Controller Load when running a GPUGrid WU?

I agree that it's better to run 2 GPUGrid WU's at a time on Einstein, but as you point out it's a CUDA 3.2 app (due to bugs) and at GPUGrid more recent CUDA versions are used (4.2, 5.5 and now 6.0).

The complexity of the program determines actual power draw and GPU usage, as does the amount of memory and the number of shaders you use... It's also the case that CPU apps use different amounts of power. I once found that a Q6600's power usage could vary by >30W depending on what Boinc apps its running (running 4 CPU apps).

When the power usage is higher you are doing more 'Work' (Applied Mechanics).

Running more that 1 WU has been exhaustively discussed. Basically, you would need a high end 3GB card and very poor performing tasks to make it worthwhile here. On my Windows system I'm getting 85 to 92% GPU usage for my GPUGrid tasks depending on the WU type. Both my GPU's are overclocked by ~10%, almost eliminating the WDDM issues.


THX for the quick answer,

- yes, I do run igpu apps, two at a time. I'm aware that this reduces memory bandwith usable for the gpu apps, but not that much, that the gpu usage is < 80%. I leave one complete CPU free, so this is usually not a problem. Anyway, it is a beta version and the final version may make better use of the gpu. Let's delay that question.
- more than one wu - a 3GB card: GPU-Z reports 300MB used ram , this should not be the bottleneck unless GPU-Z has another understanding what used ram means. Mem controller load says 22%.

I have learned that it usually makes sense to run 2 or more wu's when the gpu usage is < 80% and makes little or no sense if usage is > 90%; the overhead eats the gain then.

I want to get a deeper understanding, what gpu usage means, how accurate that value is. Without knowing more about this I tend to assume a 95% gpu load equals a 95% power load, but reality shows that this assumption is wrong.

I posted my question in the techpowerup forum, hoping that a developer is reading it and finds time to answer my question.

Anyway, thanks for the time you take to answer my questions,

cheers
Alexander

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35701 - Posted: 16 Mar 2014 | 23:54:30 UTC - in response to Message 35700.
Last modified: 16 Mar 2014 | 23:57:33 UTC

Thanks for answering my questions.

22% is quite low (and a good thing)! The larger cache size might be the reason.
I suspect that GPUGrid WU performances are being reduced by ~20% (roughly) due to running the Einstein iGPU app, but you would have to suspend the Einstein app to find out what it is for you (and this will vary depending on your CPU). I've tested this to some extent.

At GPUGrid if the apps utilize >85% of the GPU you cannot benefit from running 2 tasks. I have tested this quite extensively. If apps operate at ~80% it's not worth doing either, somewhat due to error rates which are exponential when running 2 tasks.

Different GPUGRID WU's use different amounts of RAM; ~300MB, ~600MB and over 1GB have been common over the last year or two, and when you use more than 1.5GB things get complex for most cards (a shader to Memory ratio thing; with GK104 and 2GB).
____________
FAQ's

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

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35702 - Posted: 17 Mar 2014 | 6:49:33 UTC - in response to Message 35701.



I suspect that GPUGrid WU performances are being reduced by ~20% (roughly) due to running the Einstein iGPU app, but you would have to suspend the Einstein app to find out what it is for you (and this will vary depending on your CPU). I've tested this to some extent.


We will see the difference in ~ 4hrs

Cheers

Alexander

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35706 - Posted: 17 Mar 2014 | 10:14:14 UTC - in response to Message 35702.

9.5% faster.
____________
FAQ's

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

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35714 - Posted: 17 Mar 2014 | 23:22:15 UTC - in response to Message 35706.

So what have we learnt about Maxwell, and what's the situation?

The performance is slightly below that of a GTX660, or half of a GTX Titan, but a GTX750Ti probably uses a lot less power (maybe around 45W). So it's a small to medium sized card but very efficient, and because it uses so little power it's not going to get too hot.
As it turns out, it's not bottlenecked by the stingy memory bandwidth. This could be down to the larger cache.

At present we know the CUDA6 app works on Windows for Maxwell GPU's, but those with Maxwell's are still getting CUDA 5.5 tasks, which all fail. This needs to be sorted.
Few people are going to buy a small card specifically to run short WU's so this begs the question, is the GTX750Ti fast enough to complete a Long task inside 24h? The answer is probably yes, at least for now and if you crunch 24/7.

So, who should buy a GTX570Ti?
There are 4 candidates:
People looking to replace a GTX650Ti (or lesser card),
People wanting to replace older cards (GTX400's say) with more efficient cards,
People who want to buy new entry to medium performing cards.
Those who want a desktop replacement system (laptop) - the same GPU is set to appear in the GTX800m range.
____________
FAQ's

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

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35717 - Posted: 18 Mar 2014 | 6:57:48 UTC - in response to Message 35714.



At present we know the CUDA6 app works on Windows for Maxwell GPU's, but those with Maxwell's are still getting CUDA 5.5 tasks, which all fail. This needs to be sorted.
Few people are going to buy a small card specifically to run short WU's so this begs the question, is the GTX750Ti fast enough to complete a Long task inside 24h? The answer is probably yes, at least for now and if you crunch 24/7.



I've tried to test that, but got cuda5.5 tasks only.
Answer will come later when cuda6 long runs are available.

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35720 - Posted: 18 Mar 2014 | 11:19:28 UTC

Two short questions:
- where can I see that cuda 6 wu's are available (short / long runs)? After a very long download I had to kill 3 long runs because cuda5.5 will not run on my card.
- are admins aware of the fact, that the server and the web-pages are incredibly slow?

Alexander

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,834,518,624
RAC: 292,156
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35721 - Posted: 18 Mar 2014 | 12:59:38 UTC - in response to Message 35720.

Server Status Will tell you if work is available in a queue,

Tasks by application application unsent in progress avg runtime of last 100 results in h (min-max) users in last 24h Short runs (2-3 hours on fastest card) 169 1,520 1.46 (0.04 - 9.46) 652 ACEMD beta version 0 76 1.20 (0.22 - 3.62) 93 Long runs (8-12 hours on fastest card) 21 2,017 5.26 (0.07 - 24.00) 615 CPU only app 1 70 0.15 (0.07 - 0.24) 4


Apps Will tell you which queue has a CUDA 6.0 app,

Short runs (2-3 hours on fastest card)
Platform Version Installation time

Microsoft Windows (98 or later) running on an Intel x86-compatible CPU 8.15 (cuda42) 8 Dec 2013 | 15:32:01 UTC
Microsoft Windows (98 or later) running on an Intel x86-compatible CPU 8.15 (cuda55) 8 Dec 2013 | 15:32:04 UTC
Microsoft Windows (98 or later) running on an Intel x86-compatible CPU 8.15 (cuda60) 14 Mar 2014 | 22:23:53 UTC
Linux running on an AMD x86_64 or Intel EM64T CPU 8.00 (cuda42) 26 Aug 2013 | 10:49:37 UTC
Linux running on an AMD x86_64 or Intel EM64T CPU 8.00 (cuda55) 26 Aug 2013 | 10:49:36 UTC

ACEMD beta version
Platform Version Installation time
Microsoft Windows (98 or later) running on an Intel x86-compatible CPU 8.15 (cuda42) 14 Nov 2013 | 14:36:32 UTC
Microsoft Windows (98 or later) running on an Intel x86-compatible CPU 8.15 (cuda55) 14 Nov 2013 | 14:39:01 UTC
Linux running on an AMD x86_64 or Intel EM64T CPU 8.11 (cuda60) 15 Mar 2014 | 0:07:21 UTC
Linux running on an AMD x86_64 or Intel EM64T CPU 8.14 (cuda42) 20 Sep 2013 | 21:53:36 UTC
Linux running on an AMD x86_64 or Intel EM64T CPU 8.14 (cuda55) 20 Sep 2013 | 21:53:34 UTC

Long runs (8-12 hours on fastest card)
Platform Version Installation time
Microsoft Windows (98 or later) running on an Intel x86-compatible CPU 8.15 (cuda42) 23 Jan 2014 | 14:02:55 UTC
Microsoft Windows (98 or later) running on an Intel x86-compatible CPU 8.15 (cuda55) 23 Jan 2014 | 14:02:57 UTC
Linux running on an AMD x86_64 or Intel EM64T CPU 8.03 (cuda42) 30 Aug 2013 | 7:52:50 UTC
Linux running on an AMD x86_64 or Intel EM64T CPU 8.03 (cuda55) 30 Aug 2013 | 7:52:49 UTC

CPU only app
Platform Version Installation time
Microsoft Windows (98 or later) running on an Intel x86-compatible CPU 1.05 13 Mar 2014 | 21:53:20 UTC
Linux running on an AMD x86_64 or Intel EM64T CPU 1.05 13 Mar 2014 | 21:59:13 UTC

The problem is, just because there is a CUDA6 app in the short queue doesn't mean you will get a task that will use that app - you might get tasks that use the CUDA5.5 app.

I think the server is supposed to determine your GPU and appropriate apps according to the spec, but this doesn't work. For some reason people with a GTX650Ti seem to be getting lots of CUDA6 work, but those with a GTX750Ti are not. Whatever the system, it's not working.

Is there a manual way to force tasks to use the CUDA6 app?
____________
FAQ's

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

Werkstatt
Send message
Joined: 23 May 09
Posts: 116
Credit: 122,267,748
RAC: 35,976
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35722 - Posted: 18 Mar 2014 | 13:33:30 UTC

https://dl.dropboxusercontent.com/u/50246791/gpugrid%20server%20status%20page.PNG
This is how my statuspage looks like, nothing to click on it.

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 788
Credit: 1,422,060,845
RAC: 1,410,932
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 35727 - Posted: 18 Mar 2014 | 23:21:10 UTC

I've just taken the plunge, and obtained a 750Ti: at the moment, it's in host 170387.

It may be the luck of the draw, but the first two tasks allocated have both been cuda60 (I made sure it was allowed to fetch short queue only before I attached).

The correct way to ensure this in the plan_class specification would be to add a restriction based on Compute Capability to cuda42 and cuda55: don't allocate if Compute Capability < (strictly) 5.0

Some people are saying that their hosts are trying to run earlier apps on their 750Tis. May I ask if these are in hosts with multiple GPUs - a 750(Ti) and an earlier model as well? BOINC is bad at handling this: tasks are allocated to the computer, not specifically to the card, so if you have a CC 2 or CC 3 card in the machine, it passes the test and can get cuda42 or cuda55 work - which can end up being run on the Maxwell. I'll need to think if there's a way round that.

Small side comment: it took ages to get the first task running on the new machine, while the (huge) cufft32_55 DLL downloaded. That's not needed by this app (nor is the v4.2 cufft DLL), and won't be used on this host by this project. A little optimisation of your application version generator would be welcome.

Post to thread

Message boards : News : New app on acemdbeta with Maxwell support