Advanced search

Message boards : Graphics cards (GPUs) : GTX Titan compatibility news?

Author Message
Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29107 - Posted: 10 Mar 2013 | 16:09:43 UTC

Have the scientists/developers/programmers been able to make any progress on updating the apps for GTX Titan compatibility?

I've got mine busy generating rainbow tables at present and would like to put it back to work here at GPUGrid.

Operator
____________

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29120 - Posted: 12 Mar 2013 | 1:39:44 UTC

Thought I'd add something I noticed today. The details page for my system with the GTX Titan says it has 4GB of ram.

"Coprocessors NVIDIA GeForce GTX TITAN (4095MB) driver: 314.9"

The problem is that the GTX Titan actually has 6GB of ram.

Could that account for why tasks are failing? A memory size mismatch between what the system thinks the card has and what it actually does have?

Has anybody else with a GTX Titan noticed this discrepancy?

Operator[/b][/quote]
____________

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29121 - Posted: 12 Mar 2013 | 3:07:53 UTC
Last modified: 12 Mar 2013 | 3:08:37 UTC

This is from the BOINC log at startup:

3/11/2013 5:56:31 PM | | NVIDIA GPU 0: GeForce GTX TITAN (driver version 314.9, CUDA version 5.0, compute capability 3.5, 4096MB, 8384512MB available, 4989 GFLOPS peak)

3/11/2013 5:56:31 PM | | OpenCL: NVIDIA GPU 0: GeForce GTX TITAN (driver version 314.09, device version OpenCL 1.1 CUDA, 6144MB, 8384512MB available)

Strange. Only the OpenCL section is reporting the correct memory amount.

Operator
____________

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 912
Credit: 2,216,536,145
RAC: 310,993
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29130 - Posted: 12 Mar 2013 | 19:11:24 UTC

The NVIDIA driver API seems only to return a 32-bit value.

It's been reported.

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29131 - Posted: 12 Mar 2013 | 21:11:33 UTC - in response to Message 29130.
Last modified: 12 Mar 2013 | 21:12:26 UTC

I saw over at the SETI forums that the Titan memory error was reported to the BOINC alpha list.

So I guess the BOINC devs are working that.

But are you indicating that would be a Nvidia driver problem instead?

Operator
____________

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 912
Credit: 2,216,536,145
RAC: 310,993
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29134 - Posted: 12 Mar 2013 | 23:26:41 UTC - in response to Message 29131.

Well. In the first place, those who run a 32-bit operating system will only ever see 4GB - that's the limit.

For 64-bit OSs: you'll have seen Claggy's reports of the 'wacky memory' bugs, when BOINC was reporting millions of millions of megabytes for some cards. The BOINC devs can't afford to buy a new card to test every time they come out - certainly not at Kepler and Titan prices - so Rom remoted in to my GTX 670 to work out what was going on.

Apparently, the 64-bit drivers require the programmer to pass a 64-bit memory buffer to return the memory size: but they only populate the lower 32 bits. Any initialisation garbage in the upper 32 bits remains unchanged - that's what led to the wacky memory sizes. I think Rom cured that by zeroing all 64 bits before use: if the drivers are still only filling in 32 of them, there's not much more to be done.

I wonder what the Tesla drivers do? Anyone got one they can lend to BOINC?

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29167 - Posted: 14 Mar 2013 | 14:52:21 UTC - in response to Message 29134.

I did fail to mention that I was running Win7x64. Sorry.

I have seen several event log postings over on the SETI board showing the same 4096mb memory amount (you were on that conversation as well).

I have a support ticket in with Nvidia and am in the process of a dialog with them now.

Operator


____________

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 29172 - Posted: 15 Mar 2013 | 1:47:19 UTC - in response to Message 29107.

Coming soon, once the Noelia WUs are working ok.

MJH

nucleon
Send message
Joined: 18 Dec 11
Posts: 10
Credit: 172,348,621
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwat
Message 29175 - Posted: 15 Mar 2013 | 21:29:48 UTC

Hi,

I have a titan and also I do work for other projects.

We found having the default memory speed caused too much errors.

When we clock down the memory from 3000MHz down to 2500MHz, the apps were a _lot_ more stable.

At 3000MHz, some apps would die within mins (4hrs was the best stability). Drop the speed down, stable as a rock. Ran for 48hrs+ no dramas, all results double checked aok.

Hope this helps.

-- Craig

Trotador
Send message
Joined: 25 Mar 12
Posts: 90
Credit: 1,431,740,570
RAC: 8,227
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 29176 - Posted: 15 Mar 2013 | 21:50:09 UTC - in response to Message 29175.

Hi,

I have a titan and also I do work for other projects.

We found having the default memory speed caused too much errors.

When we clock down the memory from 3000MHz down to 2500MHz, the apps were a _lot_ more stable.

At 3000MHz, some apps would die within mins (4hrs was the best stability). Drop the speed down, stable as a rock. Ran for 48hrs+ no dramas, all results double checked aok.

Hope this helps.

-- Craig


Which other distributed computing projects are Titans currently working? do they well on them?

Thanks

HA-SOFT, s.r.o.
Send message
Joined: 3 Oct 11
Posts: 100
Credit: 5,533,954,251
RAC: 115,992
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29177 - Posted: 15 Mar 2013 | 22:41:06 UTC - in response to Message 29176.
Last modified: 15 Mar 2013 | 22:43:45 UTC

Mem clocky are ok at 3000Mhz no worries.

You can check milkyway (with DP enabled) and run 5 tasks in parallel. Primegrid sieve can be ok also.

Trotador
Send message
Joined: 25 Mar 12
Posts: 90
Credit: 1,431,740,570
RAC: 8,227
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 29178 - Posted: 15 Mar 2013 | 23:25:52 UTC - in response to Message 29177.

Mem clocky are ok at 3000Mhz no worries.

You can check milkyway (with DP enabled) and run 5 tasks in parallel. Primegrid sieve can be ok also.


Thanks, I've just read your posts in Milky, 300kppd, it does not seem that impressive, does it? equal or lower than a hd 7970 if i've seen it correctly.

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2688
Credit: 1,172,901,099
RAC: 144,879
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29180 - Posted: 16 Mar 2013 | 13:23:05 UTC

@nucleon: that sounds almost like you could (should?) RMA the card. MAybe try to expose repeatable errors with some of those "artifact tester" utilities at stock clocks. Errors should also pop up there if applications crash.

@Milkyway: yeah, 300k RAC is massive for a nVidia, but not all that great. I remember scoring 260k RAC with my HD6950 maxed out (unlocked shaders, OC, 3 concurrent WUs).. back in the days before POEM came along.

MrS
____________
Scanning for our furry friends since Jan 2002

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29181 - Posted: 16 Mar 2013 | 14:10:56 UTC
Last modified: 16 Mar 2013 | 14:13:00 UTC

I now have two Titans (as does nucleon I think)and am having no difficulty running Einstein@home (GPU only tasks), as well as the DSTRGEN project.

Haven't tried PrimeGrid but good to know it works.

I have a ticket in with Nvidia on the memory reporting issue (4096mb reported vs 6144mb actual). Nothing back from that yet. The rep said that he was unaware of any issues in BOINC or GPUGrid so I pointed them toward the forums where the problems were mentioned and sent in some screenshots and they are presently "researching" the problem.

So since MW works as well, I may go back over there for a while then.. until we get the longs sorted here.

I don't think I'll be downclocking the memory just yet, unless somebody can conclusively state that doing so fixes the long work units here at GPUGrid.

Operator
____________

HA-SOFT, s.r.o.
Send message
Joined: 3 Oct 11
Posts: 100
Credit: 5,533,954,251
RAC: 115,992
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29183 - Posted: 16 Mar 2013 | 21:12:34 UTC - in response to Message 29180.
Last modified: 16 Mar 2013 | 21:12:51 UTC

Milky is open_cl so the same power as top ati is a little miracle. I have not expected that.

DistrRtGen can crunch one task in 530s with titan, overclocked titan in 480s. This is a little better then top ati.

But I'm looking forward to gpugrid. I have buy this card for gpugrid.

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29184 - Posted: 16 Mar 2013 | 22:20:35 UTC - in response to Message 29183.

But I'm looking forward to gpugrid. I have buy this card for gpugrid.



Same here my friend....

Operator
____________

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 912
Credit: 2,216,536,145
RAC: 310,993
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29188 - Posted: 17 Mar 2013 | 11:11:57 UTC

There is mention of a specific OpenCL fix for TITAN in the 314.21 beta driver release notes:

[GeForce GTX TITAN]: Application crashes or a TDR occurs when running
applications that use OpenCL. [1237211]

We don't use OpenCL here, of course - we use CUDA - but it might be worth a test.

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29204 - Posted: 20 Mar 2013 | 18:33:04 UTC

I don't know what progress is being made on GTX Titan compatibility with GPUGrid apps, but as has been stated here and in other areas the Titan seems to be happily crunching away on Einstein, Milky Way, Prime Grid, Distributed Rainbow Table generator, etc.

Comments made by Mr. Haselgrove regarding the cause of the misreporting of the total available memory (4096mb vs 6144mb) due to the driver API leaves me wondering if this is in Nvidia's lane, BOINC's lane, or both, and whether anybody is working to resolve this issue or if this issue even matters when crunching CUDA apps with a Titan.

Does this memory reporting mismatch affect how the apps run (or fail to run) on a Titan? Or is it a 'red herring'? I don't know. But obviously some projects are not having any issues at all running on the new card.

Does anyone have a timeline, idea, clue, or rumor regarding when the GTX Titan might be able to successfully support GPUGrid project apps?

Thanks,

Operator
____________

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 29205 - Posted: 20 Mar 2013 | 20:46:24 UTC - in response to Message 29204.


Does anyone have a timeline, idea, clue, or rumor regarding when the GTX Titan might be able to successfully support GPUGrid project apps?


Now we have Titans in the lab I'll be doing it late next week.

MJH

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29206 - Posted: 20 Mar 2013 | 21:08:23 UTC - in response to Message 29205.

Great news!

Thank you Sir!

Operator
____________

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 82,949
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29207 - Posted: 20 Mar 2013 | 22:29:05 UTC - in response to Message 29121.

This is from the BOINC log at startup:

3/11/2013 5:56:31 PM | | NVIDIA GPU 0: GeForce GTX TITAN (driver version 314.9, CUDA version 5.0, compute capability 3.5, 4096MB, 8384512MB available, 4989 GFLOPS peak)

3/11/2013 5:56:31 PM | | OpenCL: NVIDIA GPU 0: GeForce GTX TITAN (driver version 314.09, device version OpenCL 1.1 CUDA, 6144MB, 8384512MB available)

Strange. Only the OpenCL section is reporting the correct memory amount.

Operator


What's really strange is that the entire Kepler GeForce 600 range is actually OpenCL 1.2 capable, but the drivers only provide 1.1 functionality!

I think the Titan GPU is due a driver boost to better compete against the ATI cards in this area. Obviously the FP64 crippled 104 cards won't gain much if anything, but at least there is now a reason to support OpenCL 1.2.
____________
FAQ's

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

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29259 - Posted: 26 Mar 2013 | 23:52:55 UTC - in response to Message 29107.
Last modified: 27 Mar 2013 | 0:00:13 UTC

If you have a GeForce Titan, pending GPUGRID, the following projects are compatible,

Boinc

Collatz, Math
DistrRTgen, Math
WCG HCC GPU, Bio
Moo!, Math
Einstein, Astro
MilkyWay, Astro
Seti Beta, Astro

Folding @ Home, Bio

@+
*_*
____________

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2688
Credit: 1,172,901,099
RAC: 144,879
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29265 - Posted: 27 Mar 2013 | 16:11:42 UTC - in response to Message 29259.

BTW: Moo! is not Math, is just brute-force cracking some meaningless encrypted message.

MrS
____________
Scanning for our furry friends since Jan 2002

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29270 - Posted: 27 Mar 2013 | 23:52:13 UTC - in response to Message 29265.
Last modified: 28 Mar 2013 | 0:13:09 UTC

Encryption is usually done with primes, and the calculation of prime numbers require a lot of mathematical calculations especially if one wants to calculate prime numbers very quickly.

L'encryptage est généralement effectué avec des nombres premiers, et le calcul de nombres premiers demandent beaucoup de calculs en mathématique surtout si on veut calculer des nombres premiers très rapidement.

@+
*_*
____________

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 82,949
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29274 - Posted: 28 Mar 2013 | 19:15:47 UTC - in response to Message 29270.

All research uses maths, but using mathematics doesn't mean you're performing research!
Bottom line, it's a meaningless waste of electric and resources - a needless contribution to global warming.
It shouldn't be mentioned along side science, medicine or research.

Moo! is as batty as The Batty Bat
____________
FAQ's

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

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29286 - Posted: 30 Mar 2013 | 0:29:41 UTC - in response to Message 29274.
Last modified: 30 Mar 2013 | 0:42:28 UTC

The problem with the GeForce TITAN and GpuGrid a chance to be corrected?
I personally have tried almost all projects. But mostly I run projects that are devoted to biology and rotate with a GPU, GPUGRID, WCG HCC and Folding @ Home.
It's a shame not to be able to rotate GPUGRID.

Le problème de GpuGrid avec la GeForce TITAN à une chance d''être corrigé ?
Personnellement j'ai essayé pratiquement tous les projets. Mais je fais tourner principalement les projets qui sont consacrés à la Biologie et qui tournent avec un GPU, GPUGRID, WCG HCC et Folding@Home.
C'est dommage de ne plus pouvoir faire tourner GPUGRID.

@+
*_*
____________

Chris
Send message
Joined: 26 Nov 09
Posts: 1
Credit: 29,072,912
RAC: 0
Level
Val
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29291 - Posted: 30 Mar 2013 | 23:06:38 UTC - in response to Message 29274.

Mmmm and they dare to say that ignorance does not breed stupidity.


All research uses maths, but using mathematics doesn't mean you're performing research!
Bottom line, it's a meaningless waste of electric and resources - a needless contribution to global warming.
It shouldn't be mentioned along side science, medicine or research.

Moo! is as batty as The Batty Bat

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2688
Credit: 1,172,901,099
RAC: 144,879
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29293 - Posted: 31 Mar 2013 | 12:11:43 UTC - in response to Message 29270.

Encryption is usually done with primes, and the calculation of prime numbers require a lot of mathematical calculations especially if one wants to calculate prime numbers very quickly.

True, and that's what PrimeGrid does. Moo, on the other hand, simply uses a known and proven algorithm to find the key to a single message by chance.

@chris: you disagree with the quote from SK?

MrS
____________
Scanning for our furry friends since Jan 2002

nucleon
Send message
Joined: 18 Dec 11
Posts: 10
Credit: 172,348,621
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwat
Message 29315 - Posted: 4 Apr 2013 | 10:26:52 UTC

Touching base, I think we're now past the deadlines mentioned.

Any update?

-- Craig

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 29316 - Posted: 4 Apr 2013 | 12:49:59 UTC - in response to Message 29315.

Any update?


Not forgotten, just preëmpted.

MJH

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29350 - Posted: 6 Apr 2013 | 16:32:28 UTC - in response to Message 29316.

Any update?


Not forgotten, just preëmpted.

MJH


It wouldn't be the recent NOELIA WU's by any chance would it? ;-}

Since you folks got Titan(s) in your shop have you done any preliminary poking around to get an idea of what is making the current long WUs fail?

If so, is whatever the problem is completely fatal and requiring a complete rewrite of the program?

I've been toying with the idea of using the anonymous method to see if that would let me run long WUs.

I still don't know if the misreported memory total has anything to do with the failures but it doesn't seem to be affecting successful completion of tasks on Einstein. I have both Titans running 5 concurrent Einstein tasks each and it's running like a freight train.

Please let us Titan owners know as soon as you have something for us.

Thanks,

Operator

____________

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29351 - Posted: 6 Apr 2013 | 17:18:35 UTC - in response to Message 29350.

I have both Titans running 5 concurrent Einstein tasks each and it's running like a freight train.

Interesting, it looks like the Titan is running about 35-40% slower than an HD 7970 with 5x WUs and around the same speed as an HD 7950, according to this chart at Einstein:

http://www.dskag.at/images/Research/EinsteinGPUperformancelist.pdf

Wingless Wonder
Send message
Joined: 6 Feb 09
Posts: 3
Credit: 1,822,030
RAC: 0
Level
Ala
Scientific publications
watwatwatwatwat
Message 29352 - Posted: 6 Apr 2013 | 17:36:00 UTC - in response to Message 29351.

Interesting, it looks like the Titan is running about 35-40% slower than an HD 7970 with 5x WUs and around the same speed as an HD 7950, according to this chart at Einstein:

http://www.dskag.at/images/Research/EinsteinGPUperformancelist.pdf

I don't even see a Titan on that list...

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29354 - Posted: 6 Apr 2013 | 17:59:40 UTC - in response to Message 29352.

Interesting, it looks like the Titan is running about 35-40% slower than an HD 7970 with 5x WUs and around the same speed as an HD 7950, according to this chart at Einstein:

http://www.dskag.at/images/Research/EinsteinGPUperformancelist.pdf

I don't even see a Titan on that list...

It's not, but here's the 5x Titan results for you to peruse:

http://einstein.phys.uwm.edu/results.php?hostid=6889477&offset=0&show_names=1&state=3&appid=0

Wingless Wonder
Send message
Joined: 6 Feb 09
Posts: 3
Credit: 1,822,030
RAC: 0
Level
Ala
Scientific publications
watwatwatwatwat
Message 29355 - Posted: 6 Apr 2013 | 18:23:14 UTC - in response to Message 29354.

It's not, but here's the 5x Titan results for you to peruse:

http://einstein.phys.uwm.edu/results.php?hostid=6889477&offset=0&show_names=1&state=3&appid=0

Thanks. Hmm, interesting. I wonder where the bottleneck lies?

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29357 - Posted: 6 Apr 2013 | 20:38:20 UTC - in response to Message 29355.



This is two EVGA Titan SC (air cooled) in a Dell T3500 WS with a Xeon W3565, 12GB, Win7Pro x64, and nothing else installed that isn't necessary to run BOINC/Apps.

Admittedly, BOINC still only sees 4096mb of the 6144mb of memory on the cards. I don't know if that has anything to do with it or not.

Operator
____________

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 82,949
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29358 - Posted: 6 Apr 2013 | 20:42:26 UTC - in response to Message 29355.
Last modified: 6 Apr 2013 | 20:43:47 UTC

Thanks. Hmm, interesting. I wonder where the bottleneck lies?

BRP4cuda32nv301 - they probably need to update their apps to better utilize the titan. Their apps were built for earlier GPUs, earlier cuda versions and compilers, so they are certainly not designed for Titans.
____________
FAQ's

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

Profile Mumak
Avatar
Send message
Joined: 7 Dec 12
Posts: 92
Credit: 225,897,225
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 29359 - Posted: 6 Apr 2013 | 20:57:04 UTC - in response to Message 29358.
Last modified: 6 Apr 2013 | 21:02:08 UTC

They use CUDA 3, because their app fails with CUDA4-5. Apparently it should be a CUDA problem, that was acknowledged by nVidia, but wasn't fixed yet.

BRP4cuda32nv301 - they probably need to update their apps to better utilize the titan. Their apps were built for earlier GPUs, earlier cuda versions and compilers, so they are certainly not designed for Titans.

Wingless Wonder
Send message
Joined: 6 Feb 09
Posts: 3
Credit: 1,822,030
RAC: 0
Level
Ala
Scientific publications
watwatwatwatwat
Message 29361 - Posted: 6 Apr 2013 | 23:21:16 UTC
Last modified: 6 Apr 2013 | 23:22:11 UTC

Thank you all for the recent information and observations regarding the GTX TITAN and CUDA crunching. I chose the TITAN SC for a single-gpu solution for a double-duty gaming and distributed computing box that would use less power than my previous pair of GTX 580 SCs. So far, it hasn't disappointed. If it crunches work units faster at some future date, so much the better ;)

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2688
Credit: 1,172,901,099
RAC: 144,879
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29368 - Posted: 7 Apr 2013 | 11:15:31 UTC

I don't think the reported memory is the problem with Titan. The apps only use as much memory as necessary anyway, and 4 GB is still plenty (be it for 1 or 2 GPU-Grids or 5 Einsteins).

And regarding Titan and CUDA 3: each SMX is pretty much the same as for the other Keplers, just with some more features (more FP64 units etc.) and more registers. If these features are not being used, it should perform just as well as any other Kepler with this amount of shaders and clock speed.

MrS
____________
Scanning for our furry friends since Jan 2002

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29371 - Posted: 7 Apr 2013 | 12:02:06 UTC - in response to Message 29352.
Last modified: 7 Apr 2013 | 12:25:07 UTC

Interesting, it looks like the Titan is running about 35-40% slower than an HD 7970 with 5x WUs and around the same speed as an HD 7950, according to this chart at Einstein:

http://www.dskag.at/images/Research/EinsteinGPUperformancelist.pdf

I don't even see a Titan on that list...

Take another look, maybe you missed it ;-)
JK/Beyond

Edit: BTW, thanks to dskagcommunity at Einstein for this fine table of GPU performance!

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 815,576,358
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29374 - Posted: 7 Apr 2013 | 13:29:45 UTC - in response to Message 29371.



Edit: BTW, thanks to dskagcommunity at Einstein for this fine table of GPU performance!


No Problem ;)
____________
DSKAG Austria Research Team: http://www.research.dskag.at



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

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29376 - Posted: 7 Apr 2013 | 17:38:34 UTC - in response to Message 29368.

I don't think the reported memory is the problem with Titan. The apps only use as much memory as necessary anyway, and 4 GB is still plenty (be it for 1 or 2 GPU-Grids or 5 Einsteins).

And regarding Titan and CUDA 3: each SMX is pretty much the same as for the other Keplers, just with some more features (more FP64 units etc.) and more registers. If these features are not being used, it should perform just as well as any other Kepler with this amount of shaders and clock speed.

MrS


And yet the Titans fail to run any GPUGrid WUs still. If they're all in the Kepler family what's the problem?

Operator

____________

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2688
Credit: 1,172,901,099
RAC: 144,879
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29378 - Posted: 7 Apr 2013 | 18:25:10 UTC - in response to Message 29376.

It's software, for sure. I can't tell you more, though, since I'm not a GPU-Grid coder.

MrS
____________
Scanning for our furry friends since Jan 2002

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29429 - Posted: 12 Apr 2013 | 19:51:46 UTC - in response to Message 29374.

I have done a little performance tuning but I think there is more work to do yet.

I have reduced the workload to 4 WUs per Titan now because my CPU has only 8 threads.

I am now averaging 2600 sec. per Einstein GPU WU with an average daily of 87,453.02.

Just waiting for GPUGrid apps I can run instead.....

Operator

____________

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29437 - Posted: 13 Apr 2013 | 12:54:55 UTC - in response to Message 29429.

I have reduced the workload to 4 WUs per Titan now because my CPU has only 8 threads.
I am now averaging 2600 sec. per Einstein GPU WU with an average daily of 87,453.02.

You should post this to dskagcommunity so he can update his excellent chart:

http://www.dskag.at/images/Research/EinsteinGPUperformancelist.pdf

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 815,576,358
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29438 - Posted: 13 Apr 2013 | 14:38:54 UTC
Last modified: 13 Apr 2013 | 14:41:13 UTC

Already done, i visit the most forums from my mainprojects every day to look whats going on AND if all of my unattended machines running fine ;)

BUUUTTTT, 87k credits? Do i something wrong? My chart calculates ~66,400k

((60secs*60mins=3600secs*24h=86400secs/average calc time 2600secs)*4WUs)*500Credits per WU=~66400
____________
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: 1,995,359,260
RAC: 82,949
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29441 - Posted: 13 Apr 2013 | 15:22:37 UTC - in response to Message 29438.

Presently that Titan is averaging 67,500 credits/day:

p2030.20121214.G175.89+00.64.C.b2s0g0.00000_2144_2 157840598 13 Apr 2013 11:41:54 UTC 13 Apr 2013 15:14:43 UTC Completed and validated 2,592.23 867.46 6.05 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121227.G203.56-00.10.C.b2s0g0.00000_1712_2 156544282 13 Apr 2013 10:35:49 UTC 13 Apr 2013 14:12:40 UTC Completed and validated 2,494.34 840.74 5.87 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121219.G201.76-01.08.N.b5s0g0.00000_1960_2 157106536 13 Apr 2013 9:43:31 UTC 13 Apr 2013 13:19:40 UTC Completed and validated 2,558.20 864.25 6.03 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.18+00.24.S.b1s0g0.00000_632_0 157957373 13 Apr 2013 9:34:19 UTC 13 Apr 2013 12:52:51 UTC Completed and validated 2,549.24 852.14 5.95 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.18+00.24.S.b4s0g0.00000_2408_1 157955777 13 Apr 2013 10:56:13 UTC 13 Apr 2013 14:34:05 UTC Completed and validated 2,608.42 877.58 6.12 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.18+00.24.S.b3s0g0.00000_2040_1 157955527 13 Apr 2013 9:10:29 UTC 13 Apr 2013 12:42:25 UTC Completed and validated 2,570.75 864.14 6.03 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.30+00.48.S.b2s0g0.00000_3488_1 157954843 13 Apr 2013 10:14:54 UTC 13 Apr 2013 13:44:16 UTC Completed and validated 2,577.59 871.34 6.08 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121230.G204.18+00.61.C.b5s0g0.00000_2920_2 155731183 13 Apr 2013 9:38:29 UTC 13 Apr 2013 13:15:41 UTC Completed and validated 2,510.45 837.77 5.85 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.30+00.48.S.b5s0g0.00000_3520_1 157949136 13 Apr 2013 7:33:05 UTC 13 Apr 2013 11:04:42 UTC Completed and validated 2,511.67 848.4 5.92 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.30+00.48.S.b6s0g0.00000_3352_1 157949075 13 Apr 2013 10:02:22 UTC 13 Apr 2013 13:43:13 UTC Completed and validated 2,550.97 856.91 5.98 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.36-00.88.N.b0s0g0.00000_1496_0 157947282 13 Apr 2013 9:19:51 UTC 13 Apr 2013 12:50:45 UTC Completed and validated 2,624.66 873.09 6.09 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.54+00.94.S.b2s0g0.00000_744_2 157920940 13 Apr 2013 8:27:34 UTC 13 Apr 2013 11:58:45 UTC Completed and validated 2,561.49 870.33 6.07 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.36-00.88.N.b1s0g0.00000_1176_0 157947053 13 Apr 2013 7:23:06 UTC 13 Apr 2013 11:02:58 UTC Completed and validated 2,500.72 847.93 5.92 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.42+00.71.S.b1s0g0.00000_3424_1 157939806 13 Apr 2013 6:28:50 UTC 13 Apr 2013 10:05:35 UTC Completed and validated 2,596.96 859.72 6 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.42+00.71.S.b1s0g0.00000_1304_1 157938770 13 Apr 2013 7:59:08 UTC 13 Apr 2013 11:35:18 UTC Completed and validated 2,558.39 862.59 6.02 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121230.G204.06+00.38.S.b0s0g0.00000_80_2 155715691 13 Apr 2013 7:42:01 UTC 13 Apr 2013 11:13:55 UTC Completed and validated 2,568.63 861.53 6.01 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.42+00.71.S.b3s0g0.00000_1792_1 157935766 13 Apr 2013 7:08:56 UTC 13 Apr 2013 10:41:13 UTC Completed and validated 2,555.72 867.24 6.05 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20110111.G194.11-00.88.S.b1s0g0.00000_2352_2 155712315 13 Apr 2013 9:01:41 UTC 13 Apr 2013 12:35:12 UTC Completed and validated 2,599.02 872.62 6.09 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.42+00.71.S.b5s0g0.00000_3560_0 157932340 13 Apr 2013 5:51:03 UTC 13 Apr 2013 9:20:53 UTC Completed and validated 2,549.40 855.62 5.97 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)
p2030.20121210.G201.48-00.65.N.b3s0g0.00000_88_1 157927608 13 Apr 2013 5:02:08 UTC 13 Apr 2013 8:32:56 UTC Completed and validated 2,570.30 857.37 5.98 500 Binary Radio Pulsar Search (Arecibo) v1.33 (BRP4cuda32nv301)


Average 2,560.46
(86400/2560) *500*4=67500

____________
FAQ's

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

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29442 - Posted: 13 Apr 2013 | 15:37:04 UTC - in response to Message 29441.
Last modified: 13 Apr 2013 | 16:06:40 UTC

I didn't do any calculation I just copied and pasted the RAC on the account page.

In fact today's RAC value for ID: 6889477 is: "Recent average credit 88,240.40".

After reflection, I now realize that the RAC and the daily may not be the same.

Thanks,

Operator
____________

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2688
Credit: 1,172,901,099
RAC: 144,879
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29444 - Posted: 13 Apr 2013 | 20:47:44 UTC - in response to Message 29442.

RAC and "average credits per day" are the same. Did you include RAC from other cards or machines or CPU work?

MrS
____________
Scanning for our furry friends since Jan 2002

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 815,576,358
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29447 - Posted: 13 Apr 2013 | 21:02:34 UTC

He ran 5 WUs before, so the RAC is higher than actual until the next two weeks or so i think. Then it should be ok. We will see ^^
____________
DSKAG Austria Research Team: http://www.research.dskag.at



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

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29455 - Posted: 14 Apr 2013 | 14:21:33 UTC - in response to Message 29444.
Last modified: 14 Apr 2013 | 14:21:52 UTC

RAC and "average credits per day" are the same. Did you include RAC from other cards or machines or CPU work?

MrS


Nope. All that machine does is E@H GPU work. There are two Titans in that box.

And today's RAC for that machine/project is: 89,855.65 probably due to a little more tweaking on my part.

Just waiting for GPUGrid work.....

Operator
____________

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29456 - Posted: 14 Apr 2013 | 14:54:10 UTC - in response to Message 29455.

And today's RAC for that machine/project is: 89,855.65 probably due to a little more tweaking on my part.

RAC is a VERY poor way to calculate current production. Take your average WU time in seconds, divide it by the number of WUs run concurrently and divide the number of seconds in a day (86400) by that number. So on this case: 86400/(2600/4) and multiply that by credits/WU (in this case 500) and you get 66462 credits/day. Simple and accurate. RAC changes due to a variety of conditions and is not accurate for our purposes.

Your output for running 5x WUs is higher:

86400/(2950/5) * 500 = 73220

nucleon
Send message
Joined: 18 Dec 11
Posts: 10
Credit: 172,348,621
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwat
Message 29699 - Posted: 6 May 2013 | 4:16:09 UTC

Just bumping this thread.

Any update to Titan compatibility with gpugrid?

-- Craig

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29796 - Posted: 10 May 2013 | 15:27:13 UTC - in response to Message 29699.

Has there been any analysis of any type performed to determine why no GPUGrid apps will run on the GTX Titan?

Anything?

I was under the impression that the developers/programmers now had a GTX Titan (or two) to use in their development process.

Yet we have heard nothing at all as to what the issue(s) are.

Is there anybody successfully running a GPUGrid app using a Tesla K20 or do they have the same problem?


Will there be a date that we GTX Titan owners can expect to be able to support the GPUGrid project with our hardware?

Thank you.

Operator
____________

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 815,576,358
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29798 - Posted: 10 May 2013 | 17:59:23 UTC

In programing you cant give a date when you fix some problems, thats why so much games are such a shit today when the developer want to bring it out @ a special date ;)
____________
DSKAG Austria Research Team: http://www.research.dskag.at



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

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29810 - Posted: 10 May 2013 | 23:57:05 UTC - in response to Message 29798.
Last modified: 11 May 2013 | 0:12:20 UTC

Folding is compatible with GeForce TITAN,

For the Folding @ home

http://fah-web.stanford.edu/cgi-bin/main.py?qtype=osstats2

project and have already tried to Gromacs GPU, and they abandoned in favor of OpenMM,

https://simtk.org/home/openmm

GPUGRID uses what?

@+
*_*

Folding and Titan in action ->

https://www.dropbox.com/s/t5fs17aq678uahq/Folding.jpg
____________

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29822 - Posted: 11 May 2013 | 16:17:16 UTC - in response to Message 29810.

Zarck;

Thanks, I am now on F@H crunching away with both Titans.

Again, thanks for the tip.

Operator
____________

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29886 - Posted: 12 May 2013 | 22:48:39 UTC - in response to Message 29822.

You use amber ?

http://ambermd.org

@ +
*_*

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29908 - Posted: 13 May 2013 | 19:49:56 UTC - in response to Message 29886.

You use amber ?

http://ambermd.org

@ +
*_*


Not sure I follow. I just downloaded the F@H and am now crunching. I'm not a developer/programmer.

Operator

____________

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29909 - Posted: 13 May 2013 | 19:58:51 UTC - in response to Message 29897.

From what I gather from discussions around here Titans do work for us locally, except for the fact that a big portion of them apparently contain some kind of fault, for which reason we don't buy any more of them.


This is the first I've heard of a GTX Titan "fault".

I'm sure Nvidia would be surprised to hear that as well.

If that was true why is it that GPUGrid is pretty much the only Nvidia GPU CUDA project that the Titan cannot run (out of all the available Nvidia GPU CUDA projects)?

I have contributed to DistR Gen, Einstein, Enigma, and now F@H successfully with hardly any issues at all relating to the GPU processing of the WU using the Titan(s).

However, everytime I try to run either a long or short GPUGrid WU it fails.

Please do not now try to claim that there is a fault in every GTX Titan as the rationale behind a lack of support for this hardware. It borders on the disingenuous.

Operator



____________

Profile Mumak
Avatar
Send message
Joined: 7 Dec 12
Posts: 92
Credit: 225,897,225
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 29910 - Posted: 13 May 2013 | 20:48:35 UTC
Last modified: 13 May 2013 | 20:50:42 UTC

It might be a CUDA problem as well.
Einstein@Home for example have issues running their tasks on CUDA 4 and 5. All their attempts to run BRP fail there, so they stick with CUDA 3. Apparently nVidia has acknowledged this problem, but haven't provided a fix since a few years... CUDA 3 might be the reason why Titan runs there.
So the problem here might be a mix of acemd + CUDA 4 + GK110 hardware.

nucleon
Send message
Joined: 18 Dec 11
Posts: 10
Credit: 172,348,621
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwat
Message 29912 - Posted: 13 May 2013 | 23:48:39 UTC


In another distributed app, once we decreased the ram speed (3GHz -> 2.5GHz) everything was fine. (mersenne LL testing, FFT/memory intensive)

Problem being Titan is one of the few vid cards with ram modules on the alternate side of the board where there's no fan/cooling. Also Titan's thermal limit seems to be a lot lower than other nvidia cards. Need to keep it under 80degC (as measured via GPUz). I currently have my machine open with a floor fan pointed at it :)

I'm currently running Milkyway on mine. My dual Titan machine is currently 6th:

http://milkyway.cs.rpi.edu/milkyway/top_hosts.php?sort_by=expavg_credit&offset=0

It might make 5th.

-- Craig

Profile Mumak
Avatar
Send message
Joined: 7 Dec 12
Posts: 92
Credit: 225,897,225
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 29916 - Posted: 14 May 2013 | 4:34:38 UTC - in response to Message 29912.


I'm currently running Milkyway on mine. My dual Titan machine is currently 6th:

http://milkyway.cs.rpi.edu/milkyway/top_hosts.php?sort_by=expavg_credit&offset=0

It might make 5th.

-- Craig


Your runtime (160-230 secs) per MW task are quite high. My 7950 does it in 25-32 secs.

nucleon
Send message
Joined: 18 Dec 11
Posts: 10
Credit: 172,348,621
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwat
Message 29917 - Posted: 14 May 2013 | 4:51:35 UTC - in response to Message 29916.

How many simultaneous tasks do you do?

I do 8 on each Titan to keep GPU usage at 100%

-- Craig

Profile Mumak
Avatar
Send message
Joined: 7 Dec 12
Posts: 92
Credit: 225,897,225
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 29919 - Posted: 14 May 2013 | 5:10:09 UTC - in response to Message 29917.
Last modified: 14 May 2013 | 5:11:19 UTC

I'm running only 1 WU per GPU to maintain reasonable temperatures (summer is coming here). I tried to run 2 which was sufficient to load the GPU at 100%, but it was running too hot, so I need to improve cooling.
How long does a single WU take on Titan?

nucleon
Send message
Joined: 18 Dec 11
Posts: 10
Credit: 172,348,621
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwat
Message 29920 - Posted: 14 May 2013 | 6:05:14 UTC - in response to Message 29919.

Funny enough, 1WU takes the same amount of time as 8. @1WU GPU usage was 12.5% approx.

-- Craig

Profile Mumak
Avatar
Send message
Joined: 7 Dec 12
Posts: 92
Credit: 225,897,225
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 29921 - Posted: 14 May 2013 | 6:30:32 UTC - in response to Message 29920.
Last modified: 14 May 2013 | 6:32:47 UTC

Funny enough, 1WU takes the same amount of time as 8. @1WU GPU usage was 12.5% approx.

-- Craig


That's easy to explain - MW@H has OpenCL apps only and nVidia GPUs just perform very weak there. If they would have a CUDA app, the performance boost would be huge.

Currently my dual HD7950s do about 570K PPD on MW@H, which I think is quite close to your dual Titans.

HA-SOFT, s.r.o.
Send message
Joined: 3 Oct 11
Posts: 100
Credit: 5,533,954,251
RAC: 115,992
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29927 - Posted: 14 May 2013 | 11:40:53 UTC - in response to Message 29912.
Last modified: 14 May 2013 | 11:42:52 UTC


In another distributed app, once we decreased the ram speed (3GHz -> 2.5GHz) everything was fine. (mersenne LL testing, FFT/memory intensive)

-- Craig


I think it has nothitg to do with memory on titan. It only transfers or hides the problem.

I have done tests with genefer prime search on primegrid and the same kind of problem occurs on my 580, which was ok (moved to another pc and worked ok). And on my titan also.

But the same problem, I think, is on tesla k20 with ecc memory.

IMHO problem is somewhere between card-pci-cpu or cuda related.

I also recommend msi afterburner or similar sw to change fan speeds to higher rate to cool down card better than default.

I'm looking forward to other users opinions and experiences.

Zdenek

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29948 - Posted: 14 May 2013 | 22:59:48 UTC - in response to Message 29927.
Last modified: 14 May 2013 | 23:09:19 UTC

I do not think that will be a problem of heat or frequency, Titan automatically changes the speed according to the GPU load, if double precision mode is activated, the Titan running at 823 Mhz. Folding@home it in turns between 954 MHz and 967 MHz depending on the type of unit Folding calculated.
Under GPUGRID units go on miscalculation from the first second. I managed to calculate and validate units with my Titan for the following projects

Collatz, 
 DistrRTgen, WCG HCC GPU, 
 Moo!, 
 Einstein, 
 MilkyWay, Seti Beta,
Folding@home

Why miscalculations GPUGRID?

@+
*_*
____________

HA-SOFT, s.r.o.
Send message
Joined: 3 Oct 11
Posts: 100
Credit: 5,533,954,251
RAC: 115,992
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29954 - Posted: 15 May 2013 | 8:05:50 UTC - in response to Message 29948.

On GPUGRID there is not support for titan yet.

I'm guessing only:

1. gpugrid app is linked with cubin code, so version for CC35 cards is not included yet. In opposite to ditrrtgen which is linked with ptx code.

2. I'm not sure if CUDA4.2 supports CC35 cards (CUDA 5.0 yes but it's another story).

3. There may be other related problems with new CC (conditional compiling etc.)

Zdenek

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 82,949
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29965 - Posted: 15 May 2013 | 16:56:03 UTC - in response to Message 29954.
Last modified: 15 May 2013 | 16:56:34 UTC

In the News thread, Gianni said supporting the Titan and GTX780 will require the application to be updated,
Titan and GTX780 support

Looks like they will do this after they get hold of a GTX780, so they can facilitate both cards, which are similar, but quite different from the GTX770 (which is basically a GTX680 in a different box, and shouldn't need any app development, if the rumors are correct).
____________
FAQ's

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

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29993 - Posted: 16 May 2013 | 8:48:09 UTC - in response to Message 29965.
Last modified: 16 May 2013 | 9:00:45 UTC

I installed Boinc 7.1.1 Windows X64 for my TITAN,

http://boinc.berkeley.edu/dl/?C=M;O=D

And I received Gianni units, but there is always the same problem, GPUGRID displays error calculation from the first second calculation.

@+
*_*
____________

HA-SOFT, s.r.o.
Send message
Joined: 3 Oct 11
Posts: 100
Credit: 5,533,954,251
RAC: 115,992
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29998 - Posted: 16 May 2013 | 10:24:48 UTC - in response to Message 29993.

Again:

On GPUGRID there is not support for titan yet.

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30233 - Posted: 22 May 2013 | 22:33:51 UTC - in response to Message 29998.

TITANS is the best four?



@+
*_*
____________

flashawk
Send message
Joined: 18 Jun 12
Posts: 297
Credit: 3,550,610,997
RAC: 906,311
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 30235 - Posted: 22 May 2013 | 22:50:23 UTC

Wonder why the SLI isn't hooked up? We should find out the prices for the GTX780 and GTX770 tomorrow, if the 780 is less than $600 US, I might pick one up. I realize a lot of you guys still have kids at home and house/car payments but I don't, I guess I'm an old fart with no bills.

If I do upgrade, I'll just give away my GTX670's to anyone who wants them but only to folks here at GPUGRID.

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30236 - Posted: 22 May 2013 | 23:24:43 UTC - in response to Message 30235.

No need for SLI, the person doing this card 4 Folding calculations in the same time.



flashawk
Send message
Joined: 18 Jun 12
Posts: 297
Credit: 3,550,610,997
RAC: 906,311
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 30238 - Posted: 22 May 2013 | 23:47:40 UTC

That's very impressive. What application are we looking at here? Is that a custom skin for BOINC? With 4 Titans running full tilt in 1 machine must require a pretty beefy power supply, I wonder how many watts he's pulling at the wall?

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30240 - Posted: 23 May 2013 | 0:14:20 UTC - in response to Message 30238.

That's very impressive. What application are we looking at here? Is that a custom skin for BOINC? With 4 Titans running full tilt in 1 machine must require a pretty beefy power supply, I wonder how many watts he's pulling at the wall?

It's FAH (Folding at Home)(not BOINC). The PS is an Enermax Platimax 1200W or 1350W (8 PCIE connectors), or even could be a 1500W (10 connectors). I wonder about the cooling, don't see evidence of water cooling on the GPUs and there's no room for air flow.

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30241 - Posted: 23 May 2013 | 0:29:00 UTC - in response to Message 30240.

I wonder about the cooling, don't see evidence of water cooling on the GPUs and there's no room for air flow.


No he's strictly on air there, and that's the way it looks when they are installed.

I have two and there isn't much of a gap for airflow but it does work and seems to automatically limit itself to around 80C no matter what you're doing.

So, I'm probably going to end up water cooling mine this summer.

But first I'm going to wait for GPUgrid to come up with something for my Titans to munch before I jump into the water.

Operator

____________

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30243 - Posted: 23 May 2013 | 0:31:42 UTC

80C is a little scary.

Dylan
Send message
Joined: 16 Jul 12
Posts: 98
Credit: 366,975,962
RAC: 196,617
Level
Asp
Scientific publications
watwatwatwatwat
Message 30244 - Posted: 23 May 2013 | 0:38:52 UTC - in response to Message 30235.
Last modified: 23 May 2013 | 1:13:01 UTC

Looks like the GTX 780 is coming at about $650, according to this article. It seems like a better deal than the Titan in terms of performance offered for the price.


article:
http://videocardz.com/41960/first-geforce-gtx-780-review-leaks-out-inno3d-gtx-780-ichill-exposed

Profile Mumak
Avatar
Send message
Joined: 7 Dec 12
Posts: 92
Credit: 225,897,225
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 30247 - Posted: 23 May 2013 | 7:51:28 UTC

Yeah - cooling was the first idea that came to my mind.. With no spacing between the GPUs he cannot cool them down properly with air only. I'm wondering at what temperatures are those Titans running, but I believe they must be throttling.
Water cooling is nice, but when the ambient temperature is higher in summer, it's less efficient.

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30248 - Posted: 23 May 2013 | 8:34:55 UTC - in response to Message 30247.
Last modified: 23 May 2013 | 8:43:18 UTC

GeForce Titan an air inlet at the back of the card, the air is extracted from inside the PC outwardly.


____________

HA-SOFT, s.r.o.
Send message
Joined: 3 Oct 11
Posts: 100
Credit: 5,533,954,251
RAC: 115,992
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30249 - Posted: 23 May 2013 | 8:40:02 UTC - in response to Message 30233.

Nice. Nice.

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30250 - Posted: 23 May 2013 | 8:52:35 UTC - in response to Message 30249.


____________

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 82,949
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30253 - Posted: 23 May 2013 | 11:18:52 UTC - in response to Message 30233.
Last modified: 3 Jun 2013 | 16:21:20 UTC

TITANS is the best four?

Yes, they are the fastest single core GPU and use exhaust cooling, but the Titans are not working for here, yet - Some Beta success recently!

The GTX780 is basically a Titan, but with less cores/shaders. Probably ~15% slower for here.
The GTX780 will be much better value for money at just over half the cost (~$650), but personal preferences, space limitations... come into it for some. If its performance turns out to be 25% faster than a GTX680 (for here) that would suggest the GTX770 will be <25% faster than a GTX680.

The GTX770 (end of the month) is basically a GTX680 (GK104), but with a slight core revision, faster memory (16.6% faster) and a higher TDP. I'm curious about the performance improvements 16.6% faster memory will bring, but it's obviously at the expense of power. I would speculate that the boost might also be slightly higher and performance is at least 10% better than a GTX680, but not 20%. It will be interesting to see the price/performance of the GTX770 against the GTX780, not that I'll be splashing out on either. The performance/Watt is unlikely to favor the GTX770 (230W TDP vs 250W TDP)...

http://videocardz.com/41960/first-geforce-gtx-780-review-leaks-out-inno3d-gtx-780-ichill-exposed
____________
FAQ's

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

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30256 - Posted: 23 May 2013 | 13:54:09 UTC - in response to Message 30253.
Last modified: 23 May 2013 | 13:54:55 UTC

http://www.amazon.com/EVGA-GeForce-Dual-Link-Graphics-03G-P4-2781-KR/dp/B00CUIVTCS/ref=sr_1_17?s=electronics&ie=UTF8&qid=1369317160&sr=1-17&keywords=GTX+GeForce+780


____________

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30258 - Posted: 23 May 2013 | 14:06:03 UTC - in response to Message 30248.

GeForce Titan an air inlet at the back of the card, the air is extracted from inside the PC outwardly.


That's the way most reference cards have worked for many years. Believe me it does not alleviate the problem of running GPUs in such close proximity. I used to run some well ventilated boxes with 3 GPUs. Temps were high on the top 2 cards even at high fan speeds. I had to leave the side panel off and blow a big fan at them for cooling. Later I tried putting one of the non-shrouded MSI PE GPUs in the center position. That worked better but still hotter than I like to see. IMO there's not a whole lot of reason not to just run 2 boxes with 2 cards each. It's cooler, much cheaper and the cards will generally run faster than the 4x boxes. There is the water cooling option for the GPUs but that adds a LOT more money and you can forget the warranty on the GPUs. As far as that goes some vendors will also scrap the warranty if the card has been running too hot...

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 82,949
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30261 - Posted: 23 May 2013 | 14:29:16 UTC - in response to Message 30258.

I agree, if space isn't an issue, 2 GPU's per system is plenty for most people. Adding a 3rd makes cooling very difficult - I've always had to leave the side panel off and add extra fans. I've only had 4 small GPU's in a system once, and it was more trouble than it was worth.

While you can buy Water cooled GPU's from some manufacturers, they cost far too much.

For 3 GPU's I would water cool the CPU. This reduces ambient case temperature and heat radiation onto the top GPU. I find that a case fan aimed at the side of GPU's helps a bit, as does placing a fan externally at the rear of the case to pull the air through. I also take out all the slot cover plates.
____________
FAQ's

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

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

Scientific publications
wat
Message 30262 - Posted: 23 May 2013 | 14:42:05 UTC
Last modified: 23 May 2013 | 19:54:11 UTC

Don't take this as advertisement or anything considering that the company is founded by our lab leader. I hope I don't go against any rules here. If I do I'll just delete the post.

After trying out tons of cases and most of them failing at cooling 4 gpus correctly, our PI designed the following cases which we now use: http://www.acellera.com/products/metrocubo/
They seem to work perfectly considering the relatively simplistic design http://www.acellera.com/products/metrocubo-gpu-chassis/, so I don't know, maybe you can get similar results if you make a similar design.

Unfortunately though I think they are not sold on the market or from the company (without buying an ACEMD license, which would be probably useless for most here) :/
It's just that I find it incredibly weird that there are none (or not many?) cases available on the market that can cool 4 gpus.

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30264 - Posted: 23 May 2013 | 14:52:46 UTC - in response to Message 30262.
Last modified: 23 May 2013 | 14:54:15 UTC

Interesting link, looking around I saw:

> ACEMD is compatible with CUDA and OpenCL, the new standard framework for
> parallel and high-performance computing over different architectures
.

Any new news coming, like OpenCL support?

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

Scientific publications
wat
Message 30265 - Posted: 23 May 2013 | 14:56:53 UTC - in response to Message 30264.
Last modified: 23 May 2013 | 14:58:51 UTC

Haha yes I was shocked now that you quoted that! I asked around and it is apparently incorrect and should be fixed.
No idea if there are any plans though for OpenCL :/

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1104
Credit: 6,101,732,079
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30267 - Posted: 23 May 2013 | 16:03:34 UTC - in response to Message 30265.

Thought that might bring a smile :-)
Thanks for checking it out.

Profile Zarck
Send message
Joined: 16 Aug 08
Posts: 143
Credit: 316,354,068
RAC: 17,265
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30269 - Posted: 23 May 2013 | 16:07:55 UTC - in response to Message 30267.
Last modified: 23 May 2013 | 16:08:25 UTC

Il y a la solution de mettre une carte mère avec 7 PCI Express et de mettre une carte graphique un slot sur deux ?



There is a solution to a motherboard with seven PCI Express graphics card to a slot on both?


____________

Profile skgiven
Volunteer moderator
Project tester
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 82,949
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30272 - Posted: 23 May 2013 | 18:03:26 UTC - in response to Message 30269.

You could use PCIE riser cards, but that would require an open case/no case (rack) build. Then there is the issue of supporting 7 GPU's - 4 might be the limit on Windows, and I'm not sure how easily it can be done on Linux. Also if you're talking about high end GPU's then you would need a 1500W PSU minimum. I've seen a 13 PCIE2 slot system, so these things are doable, just very expensive.

Regarding the Metrocubo GPU Chassis. They are hand made, so they are expensive, however I wouldn't be too much of a stickler regarding the brochure info, I expect Acellera would accommodate more bespoke models including the GTX770, and presumably the GTX780 and Titan when these GPU's are compatible with the software. They might build to facilitate other PCIE based research (AMD/ATI and Intel's Xeon Phi). If you are interested contact Acellera and ask.

chassi photos
____________
FAQ's

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

matlock
Send message
Joined: 12 Dec 11
Posts: 34
Credit: 86,423,547
RAC: 0
Level
Thr
Scientific publications
watwatwatwatwatwatwatwatwat
Message 30301 - Posted: 24 May 2013 | 15:43:39 UTC - in response to Message 30235.

If I do upgrade, I'll just give away my GTX670's to anyone who wants them but only to folks here at GPUGRID.


I'll take a couple. ;) I run Linux and power cost here is 7 cents per kWh.

flashawk
Send message
Joined: 18 Jun 12
Posts: 297
Credit: 3,550,610,997
RAC: 906,311
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 30303 - Posted: 24 May 2013 | 16:09:09 UTC

No problem, I'm serious about what I said. They have to get that chip working here first so it will be a couple months at least. I know I wouldn't be able to use the water blocks that are on them (Heatkiller), so those will go with them along with the stock heatsink and fan.

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2688
Credit: 1,172,901,099
RAC: 144,879
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30308 - Posted: 24 May 2013 | 17:37:16 UTC - in response to Message 30243.

80C is a little scary.

Whatever Titan runs, you can be pretty sure it's running at (80 +/- 1)°C, unless someone changed the stock temperature target. Titan's standard behaviour is to set the boost bin (and hence power consumption) so that this temperature is reached. Improve cooling and it will still run at 80°C, just faster (until the maximum bin is reached). This temperature target can be set easily via typical OC utilities, though.

MrS
____________
Scanning for our furry friends since Jan 2002

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30849 - Posted: 14 Jun 2013 | 16:32:34 UTC

Is the GTX Titan that much different from other keplers?

Nvidia's documentation (http://www.nvidia.com/docs/IO/123576/nv-applications-catalog-lowres.pdf) says the K10, K20, K20X will run ACEMD.

So what's the deal?

Operator
____________

ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2688
Credit: 1,172,901,099
RAC: 144,879
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 30852 - Posted: 15 Jun 2013 | 11:24:15 UTC - in response to Message 30849.

Is the GTX Titan that much different from other keplers?

My take on this question is there.

MrS
____________
Scanning for our furry friends since Jan 2002

nanoprobe
Send message
Joined: 26 Feb 12
Posts: 181
Credit: 222,317,747
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwat
Message 30855 - Posted: 15 Jun 2013 | 16:57:24 UTC - in response to Message 30301.

If I do upgrade, I'll just give away my GTX670's to anyone who wants them but only to folks here at GPUGRID.

I' d be more than happy to run one here.

Post to thread

Message boards : Graphics cards (GPUs) : GTX Titan compatibility news?