Error C++ Libary

Message boards : Graphics cards (GPUs) : Error C++ Libary

Author	Message
Joe Send message Joined: 1 Sep 08 Posts: 37 Credit: 5,864,088 RAC: 0 Level Scientific publications	Message 8036 - Posted: 1 Apr 2009 \| 14:52:43 UTC
	Are the any informations about an error in the C++ Libary with acemd_6.62_windows_intelx86__cuda.exe? Win XP Prof 32Bit and Boinc 6.6.20 Kind regards Joe
	ID: 8036 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8042 - Posted: 1 Apr 2009 \| 19:49:58 UTC - in response to Message 8036. Last modified: 1 Apr 2009 \| 19:50:39 UTC
	I don't think so if you can't provide any further details. However, if you can tell us more I can already tell you that I don't know anything about this error.. MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8042 \| Rating: 0 \| rate: / Reply Quote

Joe Send message Joined: 1 Sep 08 Posts: 37 Credit: 5,864,088 RAC: 0 Level Scientific publications	Message 8054 - Posted: 2 Apr 2009 \| 7:51:10 UTC - in response to Message 8042.
	Last week I had success with many WUs. Yesterday I got a lot of compute errors with 6.6.18 and with 6.6.20. http://www.gpugrid.net/results.php?hostid=29092&offset=0&show_names=0 And two times there was a error in the MS C++ Libary with acemd_6.62_windows_intelx86__cuda.exe Don't know if it's this http://www.gpugrid.net/result.php?resultid=472595 or this http://www.gpugrid.net/result.php?resultid=478168 WU... Normaly GPUGrid 6.62.exe is between 2% till 6% of CPU time in the taskmanager. After this error the CPU time from this exe was on 50% and the system was extremly slow. After a restart I have success with two WUs and one error... There was no change in the system or in the software... Sorry, thats all I can say.
	ID: 8054 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8083 - Posted: 2 Apr 2009 \| 20:36:06 UTC - in response to Message 8054.
	That looks tough. The WUs are mostly finished by others, so it's a problem on your side. Your software should be alright, because GPU-Grid worked previously.. which points to your hardware. What you could try: - power off the machine and remove the power cord for >10 min. Odd, but I've seen this cure strange errors. - I looked through all your errors on the current result page and the 8800GT is always the last card listed. Did it break? -> Take the 8800GT out and try again - Test the 8800GT in a seperate PC, if possible. Otherwise take the 295 out for a short test of 1 WU. - Is summer approaching your region? Did the fan on the 8800GT fail or is it overly dusty? What's the temperature while crunching? - You could also try downclocking, which would help if a rise in temperature is causing the errors (it wouldn't cure it, though). MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8083 \| Rating: 0 \| rate: / Reply Quote

Joe Send message Joined: 1 Sep 08 Posts: 37 Credit: 5,864,088 RAC: 0 Level Scientific publications	Message 8113 - Posted: 3 Apr 2009 \| 9:32:39 UTC - in response to Message 8083.
	I found the error in the computer. It's the combination from my system with the GTX 295. I change the GTX295 with a GTX260 and now it works fine. http://www.gpugrid.net/result.php?resultid=479692 I test the GTX295 in Vista 64Bit computer without errors. Perhaps the newer Nvidia driver (182.50) will help me with XP and the GTX 295. I'll test it next week... Kind regards Joe
	ID: 8113 \| Rating: 0 \| rate: / Reply Quote

Michael Goetz Send message Joined: 2 Mar 09 Posts: 124 Credit: 47,698,744 RAC: 118,540 Level Scientific publications	Message 8120 - Posted: 3 Apr 2009 \| 13:09:45 UTC - in response to Message 8113.
	Hmmm... If there's a combination of computer/GPU that doesn't work right, but swapping things around makes everything work OK, perhaps it's the power supply that's the problem. Maybe the power supply in the original computer is marginal running a GTX295 but has enough juice to run the GTX260. Likewise, perhaps power supply in the second computer has enough available power for the GTX295. It could also be how the power is distributed across the different power rails in each machine, assuming the power supply uses multiple 12V rails. Just a thought. Along the same lines, the problem could be cooling. At least that one is easy to check by monitoring GPU temps. Mike ____________ Want to find one of the largest known primes? Try PrimeGrid. Or help cure disease at WCG.
	ID: 8120 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8142 - Posted: 3 Apr 2009 \| 20:29:29 UTC
	Yes, that sounds suspeciously like the power supply. Does the original PC work with the 295 if you remove the 8800GT? And error if you put the 295 and 260 in there? If "yes" then you certainly have a power problem. MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8142 \| Rating: 0 \| rate: / Reply Quote

Joe Send message Joined: 1 Sep 08 Posts: 37 Credit: 5,864,088 RAC: 0 Level Scientific publications	Message 8180 - Posted: 4 Apr 2009 \| 18:58:40 UTC - in response to Message 8142.
	On Monday I will test more... If I find something, I'll write more... Have a nice weekend and kind regards Joe
	ID: 8180 \| Rating: 0 \| rate: / Reply Quote

Joe Send message Joined: 1 Sep 08 Posts: 37 Credit: 5,864,088 RAC: 0 Level Scientific publications	Message 8244 - Posted: 6 Apr 2009 \| 13:14:14 UTC - in response to Message 8180.
	I don't belive thats a power problem. Now I put in the GTX 260 and the GTX 295. It seems to be fantastic. The first WU finished without an error http://www.gpugrid.net/result.php?resultid=498912 The both next will be finish in about one hour... If the system will be stable now, I leave both cards indside and see what happens... But where was the bug in older configuration??? Thanks for your help an kind regards Joe
	ID: 8244 \| Rating: 0 \| rate: / Reply Quote

Michael Goetz Send message Joined: 2 Mar 09 Posts: 124 Credit: 47,698,744 RAC: 118,540 Level Scientific publications	Message 8245 - Posted: 6 Apr 2009 \| 13:38:58 UTC - in response to Message 8244.
	But where was the bug in older configuration??? Diagnosing that problem requires a little more information, at least so that we can elliminate some possibilities. In the "bad" configuration, did you notice the temperature on the video card? The driver/BOIS/hardware/whatever normally keeps the temperature in a safe range, but I've seen reports of that not functioning properly. Also, what's the wattage on the power supply, and what hardware is in the system (disk drives and so forth)? Along the same lines, do you use this machine for gaming, and if so, how reliably did games work on it? If games work well, then it's unlikely it's any sort of hardware issue, since gaming is usually far more demanding than what existing BOINC GPU projects can manage. If games work, but GPUGRID doesn't, that pretty much limits it to a configuration or software problem. FWIW, most off-the-shelf computers that you would buy in a store or from a mainstream manufacturers come with power supplies that are, at best, only marginally capable of running the newest high-end video cards. Generally speaking, a GTX260 or higher card probably is capable of drawing more power than the rest of the computer components combined. Momentary spikes in high usage, and hence high power consumption, could briefly overload the power supply, causing hard to diagnose intermittent problems. If you built the computer yourself, you probably put in a big beefy power supply. If you didn't build it, chances are the power supply is, at best, in the 300 to 350 watt range, and probably a cheaply built one, too. Mike ____________ Want to find one of the largest known primes? Try PrimeGrid. Or help cure disease at WCG.
	ID: 8245 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8253 - Posted: 6 Apr 2009 \| 17:07:38 UTC
	Just to be clear: - you had the GTX 295 running along with the 8800GT, which suddenly gave you errors? - you swapped the 295 for a GTX 260 and it worked, that means you have 8800GT and GTX 260 running in the same box without problems? - now you have GTX 295 and GTX 260 running in the same box without problems? - did you try the 8800GT separately? MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8253 \| Rating: 0 \| rate: / Reply Quote

Joe Send message Joined: 1 Sep 08 Posts: 37 Credit: 5,864,088 RAC: 0 Level Scientific publications	Message 8267 - Posted: 6 Apr 2009 \| 22:47:12 UTC - in response to Message 8253. Last modified: 6 Apr 2009 \| 22:48:22 UTC
	The "bad" configuration... is XP with the GTX295 alone and the combination GTX295 and GT8800... No problems with the 260 and the GT8800 alone. AND each card ALONE under Vista works perfect And no problems for the moment in combination GTX295 with GTX260. I have three WUs with success today. http://www.gpugrid.net/results.php?hostid=29092 Also, what's the wattage on the power supply, and what hardware is in the system 500W BeQuiet, Asus P5WDG2 Prof, Intel 2,66 duo overclocked to ~3,00, one 500GB Samsung HDD and one noname DVD Rom... It's build be myself... With XP Prof 32Bit, all updates and the newest no-beta driver. Installed software: ONLY Photoshop CS3, a few tools like Everest, BOINC 6.6.20, no games, no other software like office, acrobat or something else... Just to be clear: - you had the GTX 295 running along with the 8800GT, which suddenly gave you errors? Yes - you swapped the 295 for a GTX 260 and it worked, that means you have 8800GT and GTX 260 running in the same box without problems? Just for a few days till saturday, but yes! - now you have GTX 295 and GTX 260 running in the same box without problems? Yes for today. Hope there will be no problems tomorrow ;-) - did you try the 8800GT separately? Yes, the card is ok and works. But it need much more time for finishing on WU (it was a longer time ago)... So I thought it can be me second card - for one WU in two days... Hope I answered all questions ;-) But I have another question... On XP the CPU time is very low - about 600sec. http://www.gpugrid.net/results.php?hostid=29092 With Vista I need ten times more CPU time http://www.gpugrid.net/results.php?hostid=24499 Why??? Kind regards and thanks for your help again!!! Joe
	ID: 8267 \| Rating: 0 \| rate: / Reply Quote

Michael Goetz Send message Joined: 2 Mar 09 Posts: 124 Credit: 47,698,744 RAC: 118,540 Level Scientific publications	Message 8269 - Posted: 7 Apr 2009 \| 0:16:00 UTC - in response to Message 8267. Last modified: 7 Apr 2009 \| 0:27:06 UTC
	EDIT: I just thought of an easy way to test whether the problem is the power supply. See if you can simultaneously run FurMark on all 3 GPUs (the two in the 295 plus the 8800). If you can successfully run that benchmark for an extended time on all three GPUs, then the power is OK. That benchmark really stresses the cards, and they'll draw maximum power. 500W BeQuiet, Asus P5WDG2 Prof, Intel 2,66 duo overclocked to ~3,00, one 500GB Samsung HDD and one noname DVD Rom... It's build be myself... With XP Prof 32Bit, all updates and the newest no-beta driver. Installed software: ONLY Photoshop CS3, a few tools like Everest, BOINC 6.6.20, no games, no other software like office, acrobat or something else... If I had to guess, I'd go with insufficient power as the likely cause. If I understood the configuration correctly (sorry, you've done a lot of swapping around and I may have misunderstood) the problem occurred when running both a GTX295 and a GT8800 together with a 500W power supply. Both cards work individually in that computer, but not together, correct? I looked up the spec page for a GTX 295 (http://www.evga.com/products/moreInfo.asp?pn=017-P3-1291-AR) and they recommend a minimum 680 watt power supply. That's for running a single graphics card; you're running two. If you add in the TDP of the GT8800 (110 watts), you really should be running this on an 800 watt power supply. And that assumes you're not overclocking anything. (It's also assuming you have a single 12V rail; the power supply loading is more complex if you have multiple 12v rails.) Now, those numbers are really based on running the graphics all out in a top-end game. Most of our BOINC GPU tasks don't make full use of the parallel processing abilities, and thus draw less power. Still, it's possible that you could be getting short term power peaks (short as in milliseconds) that are exceeding the amperage capacity of the power supply. When this occurs, the voltage may drop, and that could cause some errors. Obviously, this is just speculation. There are a lot of other possible causes. If it were me, I would use a larger power supply simply because they're inexpensive and it's notoriously hard to diagnose problems related to a marginal power supply. (When I got a GTX280 for my HP desktop, I replaced the stock 350W PS with a 750W unit. That's far more than I needed, but I didn't want there to be any question about the power.) Mike ____________ Want to find one of the largest known primes? Try PrimeGrid. Or help cure disease at WCG.
	ID: 8269 \| Rating: 0 \| rate: / Reply Quote

ExtraTerrestrial Apes Volunteer moderator Volunteer tester Send message Joined: 17 Aug 08 Posts: 2705 Credit: 1,311,122,549 RAC: 0 Level Scientific publications	Message 8281 - Posted: 7 Apr 2009 \| 19:22:12 UTC - in response to Message 8269.
	Well, I don't have any guesses left about what may be wrong. I don't think it's the raw power of the PSU, as the 295 runs along nicely with the 260. The only power-related thing (I can think of right now) would be a load balancing issue together with a PSU which partly failed or gets too hot now. MrS ____________ Scanning for our furry friends since Jan 2002
	ID: 8281 \| Rating: 0 \| rate: / Reply Quote

Paul D. Buck Send message Joined: 9 Jun 08 Posts: 1050 Credit: 37,321,185 RAC: 0 Level Scientific publications	Message 8295 - Posted: 8 Apr 2009 \| 5:24:48 UTC - in response to Message 8281.
	Well, I don't have any guesses left about what may be wrong. I don't think it's the raw power of the PSU, as the 295 runs along nicely with the 260. The only power-related thing (I can think of right now) would be a load balancing issue together with a PSU which partly failed or gets too hot now. If you are using HDD power and a converter plug to goto a 6 pin plug, if you have HDD on that same line, and the PSU is borderline ... when the HDD tries to spin up it can also cause fluctuations on the lines ...
	ID: 8295 \| Rating: 0 \| rate: / Reply Quote

Joe Send message Joined: 1 Sep 08 Posts: 37 Credit: 5,864,088 RAC: 0 Level Scientific publications	Message 8302 - Posted: 8 Apr 2009 \| 15:07:48 UTC - in response to Message 8295.
	Now I have 12 complete results without an error! I belive, that my system is ok - no sand in the machine ;-) The combination GTX295 and GTX260 works very well, will not be too hot (~80°C) and my 500W seems to be enough for this PC... Thanks for all your help! Kind regards Joe
	ID: 8302 \| Rating: 0 \| rate: / Reply Quote

Post to thread

Message boards : Graphics cards (GPUs) : Error C++ Libary

	About	Science	Volunteers	Performance	Forum	Join us	Donate