Message boards : Graphics cards (GPUs) : Windows app 6.62
Author | Message |
---|---|
We are now testing 6.62 which should reduce substantially CPU usage. Only few WUs will have the new app for testing. If it works we will extend it. Windows XP users please report if it works, as last time we had problems. | |
ID: 6062 | Rating: 0 | rate: / Reply Quote | |
Very good, my first WU was a 6.62 already. :D | |
ID: 6063 | Rating: 0 | rate: / Reply Quote | |
Positive here too... | |
ID: 6064 | Rating: 0 | rate: / Reply Quote | |
Roll them bad boys out!!! :) | |
ID: 6065 | Rating: 0 | rate: / Reply Quote | |
We are now testing 6.62 which should reduce substantially CPU usage. Only few WUs will have the new app for testing. If it works we will extend it. Windows XP users please report if it works, as last time we had problems. NOW THAT IS WHAT I AM TALKING ABOUT! I can't even see CPU usage on the Q9300 as that task runs ... the question will be if the task runs to completion ... at the moment I only have the one task and it was the devils own to get it to run ... but ... running it is now ... sadly it is on the 9800GT so it will be most of the day (well, the time to complete is listed as 25:33:52; internal time for tasks on this machine is 48K seconds or 13 hours) before I can report on success or failure of the task execution ... Hopefully my other machine would get one as it rummages through the stack as it would complete the task in just a few hours ... {edit}clarified execution time{/edit} ____________ | |
ID: 6066 | Rating: 0 | rate: / Reply Quote | |
Looks great! Only 2% CPU usage, can now handle one more free core for cpu project :) | |
ID: 6067 | Rating: 0 | rate: / Reply Quote | |
........... Good stuff GDF - have not returned a result as of yet; but watching them work is a pleasure - 3-5% CPU utilisation on a Quad & i7. | |
ID: 6068 | Rating: 0 | rate: / Reply Quote | |
Have not started one on the i7 though I have two queued ... the Q9300 is 12% done with only 1:14 CPU time ... and that is with the task started at about noon (well, 19:39 UTC) sure that UCT is for this post, but that will be the delta... | |
ID: 6069 | Rating: 0 | rate: / Reply Quote | |
Ok, just started the first 6.62 task on the i7 and it is acting the same as the task on the Q9300, mostly 0 CPU, with occasional flashes to 1% ... very tolerable ... | |
ID: 6070 | Rating: 0 | rate: / Reply Quote | |
Why are we still getting only v6.61 WUs? Irritating, the CPU usage is obscene... | |
ID: 6076 | Rating: 0 | rate: / Reply Quote | |
Why are we still getting only v6.61 WUs? Irritating, the CPU usage is obscene... *WE* aren't ... though mostly we are ... :) My *FIRST* 6.62 task reported as a success ... Elapsed time was 20K seconds ... I have no idea if this is a good time or not. I rummaged a bit in the list of success tasks and if it is a longer run time it is about 3,000 seconds longer which is 50 minutes ... which sounds like a lot, but, giving up a hour to each task on the GPU is worth the release of the CPU time (at least to me) if it is indeed a consequence of the change in the application and is not just an artifact of the particular model ... Without many more runs, no way to tell for sure ... {edit} With more 6.61 tasks filling up my queue I feel like a drug addict in withdrawal ... I will point out when you make the decision to change there is a thing you can update in the scheduler that will allow all currently issued work to use the new and updated application. So, when you "throw the switch" we don't have to kill current work but can do an update and work already issued will be processed with the new application version ... ____________ | |
ID: 6077 | Rating: 0 | rate: / Reply Quote | |
It's really strange now... I got some 6.62 tasks but one 6.61 on the same PC is still running...and...one of the new 6.62 task shows running but CPU and the task itself remains idle and NO computation is done for hours. I've tried to abort it, but with the next downloaded the problem remains, so from the 3 possible units (I have installed 1X9600 GT and 1X9800GX2) on quad core at the moment there are only 2 actually running..... | |
ID: 6078 | Rating: 0 | rate: / Reply Quote | |
I have aborted the pending 6.62 tasks.....and... | |
ID: 6079 | Rating: 0 | rate: / Reply Quote | |
As GDF said: Only few WUs will have the new app for testing. If it works we will extend it. I hope people with xp machines will report a go for launch! I dont like going back 6.61. BTW, does app_info.xml works here? | |
ID: 6080 | Rating: 0 | rate: / Reply Quote | |
OK the 6.62 unit itself runs OK...GPU time seems the same, low CPU usage, not like the 6.61 ones, but when mixed with the 6.61 there is some incompatibility.. | |
ID: 6081 | Rating: 0 | rate: / Reply Quote | |
Hmmm, not my experience at all... | |
ID: 6083 | Rating: 0 | rate: / Reply Quote | |
As GDF said: that will, in theory, work anywhere ... You going to try to cheat? :) ____________ | |
ID: 6084 | Rating: 0 | rate: / Reply Quote | |
GDF - works and validates very well for me; so I'd be happy to continue with 6.62 WUs - actually, it is a disappointing to be starting a new 6.61 WU this morning now the test pool has gone! | |
ID: 6085 | Rating: 0 | rate: / Reply Quote | |
With the first runs it seems to be 17% slower. On Linux the performance seems to be better. More WUs out. | |
ID: 6087 | Rating: 0 | rate: / Reply Quote | |
I just got problems with a 9800 GX2 on Server 2003 x64 with drivers 180.48. | |
ID: 6088 | Rating: 0 | rate: / Reply Quote | |
I just got problems with a 9800 GX2 on Server 2003 x64 with drivers 180.48. No, your not, I found these 4 Wu's that erred out shortly after starting it looks like on 3 different x64 Box's: http://www.gpugrid.net/result.php?resultid=256186 http://www.gpugrid.net/result.php?resultid=256301 http://www.gpugrid.net/result.php?resultid=256073 http://www.gpugrid.net/result.php?resultid=256045 I see 5 other 6.62 WU's on my Box's too, I'll try and get the Video Drivers Updated TO 180.22 before they run & see what happens ... Add another 1 to the List http://www.gpugrid.net/result.php?resultid=256055it started & erred before I could get to it & change the Drivers, not sure if the will do any good anyway. I Suspended the rest so I can have a little time to Update the V-Drivers and will start them up after I do & see if they run longer. | |
ID: 6089 | Rating: 0 | rate: / Reply Quote | |
I just got problems with a 9800 GX2 on Server 2003 x64 with drivers 180.48. Hi. Did you try 180.84 drivers (note .84, not .48)? BR, ____________ | |
ID: 6090 | Rating: 0 | rate: / Reply Quote | |
With the first runs it seems to be 17% slower. On Linux the performance seems to be better. More WUs out. I guess that is my http://www.gpugrid.net/results.php?hostid=19410]experience, because of the variability on the time for the tasks I can't say for sure. But I did two tasks, pretty easy to see which ones ... both validated ... {edit Add} for my other computer has reported and validated its first 6.62 task with another in the queue behind 2 6.61 tasks (sigh) {/edit} From my perspective, I think that 17% loss on the GPU side is an Ok cost ... lower heat among other things ... not sure about system responsiveness as I did not have three at one time to really see what it does on the i7 ... sadly I am down now to only 6.61 tasks ... Can you tell us what made such a *POSITIVE* change? If it was merely "tuning" the polling loop, it might be useful to consider my suggestion that you allow a "Performance" setting to allow people that are GPU Grid only to allow for the higher CPU use to get the maximum GPU performance out of their systems. The 6.6.1 operation as presented would be the default which I would term "Nice" and even better if you added a third option between the two of "Hybrid" which would attempt to balance the two with a loss on the CPU side but an increase on the GPU side though not as much as with "Performance" ... Personally I *LOVE* projects that allow me to choose how my systems are used so I can set things up to *MY* considerations and concerns ... ____________ | |
ID: 6091 | Rating: 0 | rate: / Reply Quote | |
Is there any recommended driver for XP64, one which surely works? | |
ID: 6092 | Rating: 0 | rate: / Reply Quote | |
From my perspective, I think that 17% loss on the GPU side is an Ok cost ... lower heat among other things ... not sure about system responsiveness as I did not have three at one time to really see what it does on the i7 ... sadly I am down now to only 6.61 tasks ... Can you tell us what made such a *POSITIVE* change? What are you Smoking this morning Paul, how can you say that 17% slower is better for the same amount of Credit & that it's a Positive thing, not in my book it isn't. We already have had our Credits slashed here and now it looks like with a longer run time for the same Credit we will have it slashed again. | |
ID: 6093 | Rating: 0 | rate: / Reply Quote | |
Is there any recommended driver for XP64, one which surely works? 180.84 doesn't work for 6.62 & Windows x64, at least not on my Box's anyway, the WU's error after 10-15 Minutes running time. The 180.84 Driver works okay with v6.61 though, I'm trying 1 now with the 181.22 Drivers, if that doesn't work I'll try & get a copy of the 180.48 Drivers & see what happens ... The 181.22 Driver seems to be working on my Box's, I have 1 6.62 running for 65 minutes now without giving an error, if anything changes on it I Post it. The 181.22 Driver doesn't have any Memory issues either for me anyway. It looks like about 1% to 2% CPU Usage, I've always ran 3&1 so when I finish a few that way I'll switch to 4&1 so I can see the difference in run times. I have a few 6.62's still that haven't started running yet so I can do that when the first ones finish for me. | |
ID: 6094 | Rating: 0 | rate: / Reply Quote | |
I just got problems with a 9800 GX2 on Server 2003 x64 with drivers 180.48. I have the 180.48 before trying to update, and since i just have 6.61, so right now i have the 180.84, but can't try the new app! Jim PROFIT | |
ID: 6095 | Rating: 0 | rate: / Reply Quote | |
From my perspective, I think that 17% loss on the GPU side is an Ok cost ... lower heat among other things ... not sure about system responsiveness as I did not have three at one time to really see what it does on the i7 ... sadly I am down now to only 6.61 tasks ... Can you tell us what made such a *POSITIVE* change? Because I am interesting in total system performance ... not GPU Grid performance over all of my other projects. So, for me the loss of 17% performance on GPU Grid is more than paid back with 21% improved performance on the i7 cores ... Were I only interested in supporting GPU Grid I would be upset at the loss of 17% performance ... which is why I suggested the triple setting in preferences ... you would go for Performance to get the most GPU Grid throughput and I would opt for Nice to get the best system performance ... different strokes for different folks ... You should know me well enough to know that I have always supported maximum possible options being made by the projects for the participants to be able to make the systems support the project the way the participant wants yet not degrade the project's ability to obtain their own goals ... In my case, I can easily make deadlines with the new application even though my RAC and throughput will drop some because my other projects will not be affected... and to compensate I will go buy a second GTX 295 meaning that because the project responded to me desires to not impact other projects, in the long run they make out too because I get my cake and i can eat it too ... another project supported, higher numbers of tasks done and yet GPU Grid is not misusing my systems (at least not to my eyes... sorry, I feel a polling loop that is just doing idle checking of the GPU is a poor investment) ... So, yes, GPU Grid losses some efficiency and processing speed and yet, I think, with lower CPU costs they will in the long run stand to gain more ... heck their load on the CPU is less than SaH ... and that would incline me to stay here ... Besides, though I am as addicted to my credit score as the next fanatic (well, maybe not as much as some) for me it is about getting the science done and the credit is only the reflection of the fact that I have done the work... and as such, highest RAC/credit is not what I am about ... if it were, I would only be supporting those projects that pay the most ... and I would ignore the rest ... yet my signature shows that I am about supporting ALL science and to a lesser amount all projects ... Anyway, that is why I proposed that the project consider the compromise application so that you and I can both be happy here ... if they go back to only a high performance option then I will only be here for the time it takes for a project like Einstein to come along with a GPU application (or Milky Way if you have seen the news, though the first cut looks to be a version of ATI cards) and then I will leave if other projects offer lower CPU loads ... in any case, if the load goes back up to levels I consider rediculous I will certainly not be adding more GPUs until I can use them and my CPUs efficiently ... Just my opinion, long winded as usual ... :) ____________ | |
ID: 6098 | Rating: 0 | rate: / Reply Quote | |
No go here also. XPpro 64, GTX260 6.5.0/180.84 | |
ID: 6103 | Rating: 0 | rate: / Reply Quote | |
No go here also. XPpro 64, GTX260 6.5.0/180.84 Try the 181.22 if they will install for you, I have Two 6.62's running @over 50% done now. The 181.22's installed on 9 XP Pro 64-Bit Box's of mine with no problems at all so they should for you too ... | |
ID: 6105 | Rating: 0 | rate: / Reply Quote | |
Because I am interesting in total system performance ... not GPU Grid performance over all of my other projects. So, for me the loss of 17% performance on GPU Grid is more than paid back with 21% improved performance on the i7 cores ... Maybe GDF can slow them down another 88% & that would give you 88% more Performance for your Precious i7 ... :P ... The Option is already there if you want more Performance for your GPU WU's or the Regular WU's, just run 3 & 1 for GPU Performance or 4 & 1 for Regular WU Performance so there's no need or reason to make them run longer just so they use less CPU% ... | |
ID: 6106 | Rating: 0 | rate: / Reply Quote | |
First 6.62-WU finished (257219), unusual slow 47,7 ms (GTX 260/216), but maybe GPU usage overlapped with crunching a WU for folding@home. Windows XP users please report if it works, as last time we had problems. I have an eye on it, three WUs (hostID 23101, GTX 260/192) are in queue, calculation time with manager 6.4.5 seems to be (too) long (22h, 25h, 49h). | |
ID: 6107 | Rating: 0 | rate: / Reply Quote | |
Because I am interesting in total system performance ... not GPU Grid performance over all of my other projects. So, for me the loss of 17% performance on GPU Grid is more than paid back with 21% improved performance on the i7 cores ... Hate to accuse you of not reading my posts, but, you didn't read it carefully. I feel that a 21% CPU load is too high and the exchange of 21% drop in CPU load for a 17% drop in GPU performance is a good trade-off for *ME* because I now have only a 1% CPU load per core on both my systems instead of a complete waste of 21% for 3 cores on the i7 to support 3 GPU cores and 22% on the Q9300 to support only one core. The *IDEA* of GPU processing is that it would be in parallel with CPU processing with minimal impact on normal system processing. The old GPU application had a very high load, one which my memory says that you objected to also ... We can debate the consequences of the change and how we would prefer the system to operate and again, this is the point of having OPTIONS ... were I desireous of maximum GPU Grid performance I would also agree with you that this is a bad change, or at least I am interpreting what you are saying is that this is a bad change ... but, I think that this is a good change. Again, were we to have the option you could run the application as 6.61 where you have the 17% GPU performance because you don't mind the impact on the CPU side ... and I would be able to run it as 6.62 because I do not think that the performance loss is necessacerily that bad. Especially because it is entirely possible that they can take another step forward and still keeping the CPU load low that they can regain part or all of that performance loss ... ____________ | |
ID: 6108 | Rating: 0 | rate: / Reply Quote | |
On another note I had a weird screen come up and it is unclear if it is my LCD dying or the GPUs going out to lunch, or something completely unrelated ... scary though ... | |
ID: 6109 | Rating: 0 | rate: / Reply Quote | |
I feel that a 21% CPU load is too high and the exchange of 21% drop in CPU load for a 17% drop in GPU performance is a good trade-off for *ME* because I now have only a 1% CPU load per core on both my systems instead of a complete waste of 21% for 3 cores on the i7 to support 3 GPU cores and 22% on the Q9300 to support only one core. Run in Linux then & you don't get any Performance loss, Oh thats right you don't do Linux :P hahahaha J/K ... ;) | |
ID: 6110 | Rating: 0 | rate: / Reply Quote | |
I am having a weird problem when running 6.62 apps- the other project running as 4+1 will not fetch new work until the very last of it's WU's has completed... then it fetches the max number possible and then runs them all down to zero and repeats...this happens on all quads (XP Home)... as soon as a 6.61 WU runs the other project fetches work normally... I run the same preferences on all machines- network connection set to '0' and .5 day cache... | |
ID: 6111 | Rating: 0 | rate: / Reply Quote | |
The *IDEA* of GPU processing is that it would be in parallel with CPU processing with minimal impact on normal system processing. If it reduces CPU load to 1% but takes 17% longer to run, I'd rather stick with 6.61. My GPU's crank out way more work than my CPU's - so I say screw one core and run the GPU's fast until we can find a fix that doesn't slow them down. Paul, any chance you can reduce the size of your sig? It's a PITA when trying to read threads cause you post it every time. | |
ID: 6112 | Rating: 0 | rate: / Reply Quote | |
No go here also. XPpro 64, GTX260 6.5.0/180.84 Heck, I thought 180.84 was the sweet driver (for XP64) until this. Will try 181.22 or newer. BR, ____________ | |
ID: 6113 | Rating: 0 | rate: / Reply Quote | |
some shredded 6.62 WUs here: | |
ID: 6114 | Rating: 0 | rate: / Reply Quote | |
Paul, any chance you can reduce the size of your sig? It's a PITA when trying to read threads cause you post it every time. Don't get your hopes up. He just about had an aneurysm trying to get someone to fix the site so he could post that sig... ____________ Consciousness: That annoying time between naps...... Experience is a wonderful thing: it enables you to recognize a mistake every time you repeat it. | |
ID: 6115 | Rating: 0 | rate: / Reply Quote | |
Please report timing here from 6.61 to 6.62 possible on same WU name. So that we can better estimate performance. | |
ID: 6116 | Rating: 0 | rate: / Reply Quote | |
Paul, any chance you can reduce the size of your sig? It's a PITA when trying to read threads cause you post it every time. Just go into your Account/Message boards and private messages/Hide signatures & put a check there and you won't see Signatures ... | |
ID: 6117 | Rating: 0 | rate: / Reply Quote | |
Please report timing here from 6.61 to 6.62 possible on same WU name. So that we can better estimate performance. I finished a WU HERE using v181.22 V-Card Driver & 6.3.21 BOINC Client on a Win XP Pro x64 Box & have several other almost done with the same setups. The Work unit took about 8-10 Minutes longer than Normal for that Type of WU running with 3 & 1 Settings. I'm going to switch the Box to 4 & 1 & run another 6.62 on it to see if there's a difference & how much. Will report on it later this afternoon after it finishes. | |
ID: 6118 | Rating: 0 | rate: / Reply Quote | |
Got 1 error.
| |
ID: 6119 | Rating: 0 | rate: / Reply Quote | |
Ohh, 6.62 is so sweet thing! Thanks to all developers!! 4-6% load on one core! At last we can crunch with pleasure! Even while gaming... | |
ID: 6120 | Rating: 0 | rate: / Reply Quote | |
Ohh, 6.62 is so sweet thing! Thanks to all developers!! 4-6% load on one core! At last we can crunch with pleasure! Even while gaming... Yeah but whats your Gaming going to be like if the GPU is Crunching WU's ... ??? | |
ID: 6121 | Rating: 0 | rate: / Reply Quote | |
Just go into your Account/Message boards and private messages/Hide signatures & put a check there and you won't see Signatures ... Yep, I've never seen Pauls sig or anyone elses :) | |
ID: 6122 | Rating: 0 | rate: / Reply Quote | |
Please report timing here from 6.61 to 6.62 possible on same WU name. So that we can better estimate performance. Last 6.61 WU (an SH2 type) on my machine with the GTX260 (192 shader) card took 6 hours 36 minutes. First 6.62 WU - also an SH2 type has reached 35.5% in 2 hours 25 minutes. Extroplating that indicates an elapsed time of 6 hours 48 minutes is likely. One key difference with 6.62 - I am back running in 1 plus 4 mode on this machine with just a 12 minute time penalty per GPU task compared to version 6.61. Both runs are on Vista HP 32 bit, Boinc 6.5.0 and Nvidia drivers 178.24. No games on the machine but Firefox surfing and Outlook e-mails are still responsive unlike when I tried 6.61 in 1 plus 4 mode. I've just got to wait now until my machine with the slower card completes its last 6.61 WU in about 15 hours. Phoneman1 | |
ID: 6123 | Rating: 0 | rate: / Reply Quote | |
Ah... much better. PoorBoy: thank you. | |
ID: 6124 | Rating: 0 | rate: / Reply Quote | |
Please report timing here from 6.61 to 6.62 possible on same WU name. So that we can better estimate performance. Sorry, have not the same WU name. First 6.62-WU with WinXP/32, nVidia 181.22: 241966 - 2478 credits - GTX 260/192 @666/1512/1150 - CPU time: 366.7031 (3+1 configuration) - Time per step: 43.210 ms - Approximate elapsed time for entire WU: 21605.234 s "Old" 6.61-WU with WinXP/32, nVidia 181.20 (think so): 243068 - 2478 credits - GTX 260/192 @666/1512/1150 - CPU time: 9720.516 (3+1 configuration) - Time per step: 35.068 ms - Approximate elapsed time for entire WU: 17534.094 s | |
ID: 6127 | Rating: 0 | rate: / Reply Quote | |
No go here also. XPpro 64, GTX260 6.5.0/180.84 That did the trick, seems to work fine with seti also. Thanks PoorBoy Fish | |
ID: 6128 | Rating: 0 | rate: / Reply Quote | |
Run in Linux then & you don't get any Performance loss, Oh thats right you don't do Linux :P hahahaha J/K ... ;) Actually I do have one Linux system, and I could not get any work on that system with two different GPU cards ... When I have had never had problems keeping the WIndows XP systems with work (since the server has been fixed) I decided to stay on Windows for now. When I hear what are the project plans for 6.62 for sure I am leaning towards getting another 295 card to put into the i7 and that will let me bump cards down so I can put the 9800 GT into the Linux box ... at that point I can fiddle with it for a couple days to see if I can get the Linux system to download work and keep busy ... ____________ | |
ID: 6130 | Rating: 0 | rate: / Reply Quote | |
Well, my first 6.62 has just completed 3 minutes quicker than I had calculated when I posted earlier in this thread. | |
ID: 6131 | Rating: 0 | rate: / Reply Quote | |
Question to the developers: What was the 6.61 code doing to use 5.5 times more CPU time for the same job type???? A poll loop ... All it did was to to check to see if the GPU needed the next block of data. Because of the timing of the OS, and it gets deep quickly, one of the ways to detect an idle GPU is to ask it if it is done; loop, rinse, repeat ... as fast as the CPU can make the test ... When the program loosens the timing on the CPU it is possible that the GPU goes idle for a short time. This is why the GPU application takes a little longer running on the GPU with the decreased load on the CPU ... This is the argument on what is the "proper" organization and CPU load ... ____________ | |
ID: 6133 | Rating: 0 | rate: / Reply Quote | |
Question to the developers: What was the 6.61 code doing to use 5.5 times more CPU time for the same job type? The polling loop, actively asking the GPU "are you finished yet?", like almost every other previous client. The more interesting question would be "How is 6.62 doing it?" MrS ____________ Scanning for our furry friends since Jan 2002 | |
ID: 6135 | Rating: 0 | rate: / Reply Quote | |
Question to the developers: What was the 6.61 code doing to use 5.5 times more CPU time for the same job type? Which is a question I asked ... :) | |
ID: 6137 | Rating: 0 | rate: / Reply Quote | |
On one of my XP machines (my first 6.62 WU): | |
ID: 6148 | Rating: 0 | rate: / Reply Quote | |
Well, one of 6.62 WUs has been finished already... | |
ID: 6152 | Rating: 0 | rate: / Reply Quote | |
Running 6.62 on a Q9550 w/ 2xGTX 260 Core 216's superclocked, WIN XP Pro 32bit, Nvidia drivers 180.48 successfully and getting CPU times +/- 3-4 mins per WU. Since upgrading to BOINC 6.6.3 earlier this evening I've been monitoring this box with taskmgr and the CPU utilization barely registers on these WU's. The WCC tasks are humming along nicely at 24-25% each... Amazing. | |
ID: 6154 | Rating: 0 | rate: / Reply Quote | |
No go here also. XPpro 64, GTX260 6.5.0/180.84 Same here. Thanx PB! BR, ____________ | |
ID: 6159 | Rating: 0 | rate: / Reply Quote | |
Good to hear, a Teammate of mine posted in our Teams Forum about error's with the 6.62 with 64 Bit Windows right on top of a Post I made about Updating their Video Drivers to 181.22, I'll let him figure it out on his own if he can't read whats Posted ... ;) | |
ID: 6160 | Rating: 0 | rate: / Reply Quote | |
I am not sure that I am really seeing a decrease in productivity. Looking at 14 tasks run on my computer W2 which have been run on my GTX 295 and GTX 280 ... all the tasks seem to be of the same type. GPU Total TIme Step (ms) version Task ID 295 32,796 43.72 6.62 259000 295 17,102 34.2 6.61 258856 280 15,996 31 6.61 258699 295 25,243 33.65 6.61 258509 295 17,277 34.55 6.61 257990 280 16,063 32.127 6.61 257975 295 17,339 34.67 6.61 257861 280 17,128 34.258 6.62 257485 295 18,088 36.177 6.62 257364 295 17,624 35.249 6.62 257038 295 17,493 34.982 6.62 257007 295 18,396 36.793 6.62 256746 295 18,401 36.803 6.62 256602 280 17,120 34.24 6.62 256039 As I look at these numbers, granted not a HUGE sample, but, 50 / 50 roughly there is only a slight increase in step size and a minor increase in run time ... I make it 18,170 sec average (5 hours) and 33.36 ms for the 6 each 6.61 tasks in the list... and 19,630 sec average (5 hours 27 Minutes) and 36.52 ms for the 8 each 6.62 tasks ... which is a change of roughly 8-9% ... The one other point that I would raise is that there is sufficient variation in the runtimes and time steps that the actual LONG TERM values may even be lower ... but the tedium of gathering the data means that I am not sure that there is sufficient reason to dig deeper. Also not clear is the impact on the actual daily average credit. I have only been doing GPU Grid for a little over a month and the numbers on the stat sites don't seem to align with what I am seeing in my account, though I have not tried to do my own data capture... worse, I have been moving cards around and increasing the number I have which also makes my account a poor candidate to establish a baseline this way ... If the change impacts credit grant I would expect my daily current average of about 43.5K per day would drop to about 40K ... Another caveat is that I have two different classes of card doing the processing ... anyway, make of it what you like ... I am seeing a very small increase in GPU time and a huge decrease in CPU usage ... I am not sure for the credit hounds if the one balances the other ... but for those of us that are supporting multiple projects I still think that this is a positive development ... {edit} I was getting about 4.8 tasks done per day per core with a total of 14.4 total per day (6.61) With the 6.62 version and an increase of 30 min per task my calculations indicate this drops to 12.9 total per day or a reduction of 1.5 tasks per day ... Meaning if I exchange the 280 for another GTX 295 I will still see a significant increase in processing... but that is just me I suppose .... | |
ID: 6161 | Rating: 0 | rate: / Reply Quote | |
After 6.61 I was about to hang up on GPUGRID but with the 6.62 I'm back on board full force. | |
ID: 6166 | Rating: 0 | rate: / Reply Quote | |
Excellent results here on the first 6.62 WU. Has made my system much more responsive for the additional Boinc clients I run. | |
ID: 6167 | Rating: 0 | rate: / Reply Quote | |
Sorry, have not the same WU name. Next two 6.62-WUs at the same computer (WindowsXP/32, 181.22) 239626 - 2478 credits: - GTX 260/192 @666/1512/1150 - CPU time: 366.4062 - Time per step: 36.637 ms - Approximate elapsed time for entire WU: 18318.438 s 258100 - 2478 credits: - GTX 260/192 @666/1512/1150 - CPU time: 545.2812 (switched back meanwhile to 4+1 configuration using manager 6.5.0) - Time per step: 38.747 ms - Approximate elapsed time for entire WU: 19373.393 s | |
ID: 6170 | Rating: 0 | rate: / Reply Quote | |
v6.62 is actually a bit faster here: My last v6.61 WU: CPU time 47005.82 # Using CUDA device 0 # Device 0: "GeForce 9600 GSO" # Clock rate: 1674000 kilohertz # Total amount of global memory: 402325504 bytes # Number of multiprocessors: 12 # Number of cores: 96 # Time per step: 110.087 ms # Approximate elapsed time for entire WU: 55043.295 s Validate state Valid Claimed credit 2478.98611111111 Granted credit 2478.98611111111 application version 6.61 My first v6.62 WU: CPU time 546.0781 # Using CUDA device 0 # Device 0: "GeForce 9600 GSO" # Clock rate: 1674000 kilohertz # Total amount of global memory: 402325504 bytes # Number of multiprocessors: 12 # Number of cores: 96 # Time per step: 108.343 ms # Approximate elapsed time for entire WU: 54171.699 s Validate state Valid Claimed credit 2478.98611111111 Granted credit 2478.98611111111 application version 6.62 CPU time was cut from 47005 to 546, what an improvement! | |
ID: 6172 | Rating: 0 | rate: / Reply Quote | |
I confirm 6.62 working fine on XP 32-bit with 8800GT. CPU usage usually below 2% on a Athlon64x2 4400+. I'm happy! | |
ID: 6193 | Rating: 0 | rate: / Reply Quote | |
No edit button? | |
ID: 6194 | Rating: 0 | rate: / Reply Quote | |
No edit button? You can only edit a post for up to an hour after posting... then no more edits ... have to make a new post ... But the sentiment is correct ... great work ... my daily number went down by about the amount I would have expected. Tomorrows will be lower as I did an experiment with SETI@Home Beta with CUDA and have had a new experience with the world of bugs ... The good news is that I ordered another GTX 295 this morning ... should be here mid week next and i can get my number back up ... | |
ID: 6195 | Rating: 0 | rate: / Reply Quote | |
Here my values for the 6.62 on Vista 64: | |
ID: 6198 | Rating: 0 | rate: / Reply Quote | |
Finished one on on 6.62 XP64 bit, 181.22 Nvidia driver. All good. | |
ID: 6199 | Rating: 0 | rate: / Reply Quote | |
Finished one on on 6.62 XP64 bit, 181.22 Nvidia driver. All good. It appears that there never was a problem with v5.56 in XP64 either, just a bug in the NVidia drivers. | |
ID: 6218 | Rating: 0 | rate: / Reply Quote | |
v6.62 is actually a bit faster here: I think I'm finding the same thing. I don't see things slowing down. I compared several of my 6.62 "appx elapsed times" and see they dropped about 1500-2000 seconds. | |
ID: 6219 | Rating: 0 | rate: / Reply Quote | |
v6.62 is actually a bit faster here: The problem is that there is variability in the total execution times of the tasks apparently depending on the complexity of the model or other factors which may be the interaction of the tasks with the system and OS ... This is why I looked at 14 tasks and calculated an average between the 6.61 and 6.62 versions. To be more fair I should have sampled more, but, I had not done that many 6.62 tasks by that time so I would have had problems coming up with enough tasks done by that application... | |
ID: 6224 | Rating: 0 | rate: / Reply Quote | |
Just small info: Also, let us know if any Win Xp driver works. I know that`s pretty late, but good news are always good news I hope. 181.22 works absolutelly perfect for WinXP Pro SP2 and SP3. I`ve already crunched many WU`s with this driver. About Win x64 I can just write that I don`t see anything better than 178.24. I know this driver is not perfect, but was best for me until now. ____________ | |
ID: 6228 | Rating: 0 | rate: / Reply Quote | |
Have anybody seen a difference of WU-time from 19000s - 30000s on a GTX260 with the same credit rate (4+1)? Normally Iam loosing around 2000s but this is mystic. | |
ID: 6242 | Rating: 0 | rate: / Reply Quote | |
Let's see if I have any useful performance data:
All are run in 4+1 mode. I'm seeing fairly consistent times and an average of 70.13 ms/step with 6.62 versus 68.08 ms/step with 6.61. That's a performance loss of 3%, or 2.2% if the 71.756 ms value from 6.62 is omitted (may have been playing Civ 4). System responsiveness ahs not been terrible before and is not great now, but it's certainly improved. Overall I appreciate the new client.. for me a 3% loss on the GPU is more than made up for by the benefits of the new client. And an interesting question: which part of the WU name actually tells me if they're the same? Judging from the runtime my WUs were surely similar enough of a comparison. Edit: still using 178.24 on XP32, no problems. MrS ____________ Scanning for our furry friends since Jan 2002 | |
ID: 6254 | Rating: 0 | rate: / Reply Quote | |
All are run in 4+1 mode. I'm seeing fairly consistent times and an average of 70.13 ms/step with 6.62 versus 68.08 ms/step with 6.61. That's a performance loss of 3%, or 2.2% if the 71.756 ms value from 6.62 is omitted (may have been playing Civ 4). My data showed a slightly larger loss of as much as 9% ... yet my daily numbers don't reflect that ... so, I am not sure what the reality is ... The problem as I see it is that we have slight instability in the run times of the tasks to begin with and that makes it hard to really pin down the "true" performance. Of course the other problem is that you also get the issue where one task or more falls over the boundary so that you have as more tasks done on some particular days than others. Even worse for my account is that I have been changing the number and class of the GPUs I have on hand so that the numbers keep shifting ... and will change again this week ... :) Interesting times ... | |
ID: 6260 | Rating: 0 | rate: / Reply Quote | |
Compared to previous versions / performance investigations I find these runtimes to be surprisingly consistent. But maybe that's just because I don't have time to actually use my machine over the week.. :p | |
ID: 6263 | Rating: 0 | rate: / Reply Quote | |
Another one with doubled time + same credits, I cant explain me why ^^ | |
ID: 6271 | Rating: 0 | rate: / Reply Quote | |
6.62 application fails on my 9600GT :( | |
ID: 6276 | Rating: 0 | rate: / Reply Quote | |
Another one with doubled time + same credits, I cant explain me why ^^ Well, the credit awards are fixed. There are three sizes ... so, regardless of how long it takes you are going to be granted one of the three amounts of credit. | |
ID: 6279 | Rating: 0 | rate: / Reply Quote | |
I noticed this wu hung for 12hrs or so last night at around 61.2%. A stop and start of boinc got it going again. Can't give any more info than that. | |
ID: 6281 | Rating: 0 | rate: / Reply Quote | |
I noticed this wu hung for 12hrs or so last night at around 61.2%. A stop and start of boinc got it going again. Can't give any more info than that. Ok, but this WU after the hang on and restart finished generating valid result...mine are giving only errors :( I'm already suspecting hardware failure.. | |
ID: 6289 | Rating: 0 | rate: / Reply Quote | |
WOW! look at the reduction in CPU time here! Running stably on Boinc 6.6.3 and Win App 6.62...... And for the first time for days it's reliably running in 2+1 mode not 1+1, fingers crossed it continues.....
| |
ID: 6291 | Rating: 0 | rate: / Reply Quote | |
I've seen a similar drop in CPU times. | |
ID: 6293 | Rating: 0 | rate: / Reply Quote | |
I noticed this wu hung for 12hrs or so last night at around 61.2%. A stop and start of boinc got it going again. Can't give any more info than that. FYI Up to now I had two WUs that didn't start although status has been shown as 'running'. Restarting BOINC helped in this case, too. | |
ID: 6299 | Rating: 0 | rate: / Reply Quote | |
NICE improvement | |
ID: 6323 | Rating: 0 | rate: / Reply Quote | |
NICE improvement With the 6.62 application you should not have to reserve a core to feed the GPU. I run 3 GPU cores on my i7 and have full speed on 8 tasks on the CPUs ... | |
ID: 6329 | Rating: 0 | rate: / Reply Quote | |
Running stably on Boinc 6.6.3 and Win App 6.62...... And for the first time for days it's reliably running in 2+1 mode not 1+1, fingers crossed it continues..... My problems started with 6.62 wu's. Now my computer is running in 1+1 mode all the time as earlier I was mostly able to get it to 2+1. My boinc version is 6.4.5 though | |
ID: 6354 | Rating: 0 | rate: / Reply Quote | |
My problems started with 6.62 wu's. Now my computer is running in 1+1 mode all the time as earlier I was mostly able to get it to 2+1. My boinc version is 6.4.5 though Try 6.5.0. I noticed this wu hung for 12hrs or so last night at around 61.2%. A stop and start of boinc got it going again. Can't give any more info than that. I also had a 6.62 WU hanging and restarting BOINC solved the problem. Didn't have anything like that before. (which may be coincidence) IIA2000 wrote: BOINC recognizes the card, assigns task that starts and then stays at 0% done and 0 time computing.... IIA2000 wrote: Ok, but this WU after the hang on and restart finished generating valid result...mine are giving only errors :( do they error out or are they stuck and you abort them? How long have you let them run to see if they're really stuck? With 6.62 you won't see the cpu time count up, so you can not readily tell if anything is running. Looking at the GPU temperature would tell you, though. Or waiting >10 min should also make sure that you see progress, even on the slowest of cards. And which driver are you running, 181.22? Older ones are causing problems / errors on XP64. MrS ____________ Scanning for our furry friends since Jan 2002 | |
ID: 6365 | Rating: 0 | rate: / Reply Quote | |
My problems started with 6.62 wu's. Now my computer is running in 1+1 mode all the time as earlier I was mostly able to get it to 2+1. My boinc version is 6.4.5 though That did the job. Thanks. | |
ID: 6380 | Rating: 0 | rate: / Reply Quote | |
My problems started with 6.62 wu's. Now my computer is running in 1+1 mode all the time as earlier I was mostly able to get it to 2+1. My boinc version is 6.4.5 though First of all the problem has gone...now I have 3 simultaneous running tasks as earlier. Again latest drivers and BOINC 6.6.4 As I mentioned in my post I had clean install of the latest drivers and latest CUDA...and I had no problems with the earlier versions...so I think it was not an issue with the drivers. I am running XPx64. The problem units really stuck for an hours...It's not of waiting 10-15 minutes to see the progress.. About the temperatures....YES I've mentioned that the GF9600GT card was idle...at the same moment GF9800GX2 was running 2 tasks. Here are again links to some of the problem WU: http://www.gpugrid.net/result.php?resultid=268000 http://www.gpugrid.net/result.php?resultid=267467 http://www.gpugrid.net/result.php?resultid=267302 ____________ | |
ID: 6381 | Rating: 0 | rate: / Reply Quote | |
Message boards : Graphics cards (GPUs) : Windows app 6.62