Message boards :
Number crunching :
new workunit queue size (6)
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next
Author | Message |
---|---|
Send message Joined: 15 Jan 09 Posts: 169 Credit: 6,734,481 RAC: 0 |
I don't know how many GPUs are already in use on this project, but with more likely to be used soon I would think it a good idea to increase the size of each wu, so there will not need to be so many requests. Larger work units would be a real fix, and would give you more time to sort out the problem with how slowly work is being released at times. |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
I don't know how many GPUs are already in use on this project, but with more likely to be used soon I would think it a good idea to increase the size of each wu, so there will not need to be so many requests. I brought this up a couple weeks ago and Travis said they contain enough info and the 'group' doesn't want to increase the size any more. Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 15 Jan 09 Posts: 169 Credit: 6,734,481 RAC: 0 |
I don't know how many GPUs are already in use on this project, but with more likely to be used soon I would think it a good idea to increase the size of each wu, so there will not need to be so many requests. LOL, I'd love to know what majority group thinks everything is fine as it is :) |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
Whoever he is supposed to see about the project. Probably the same people who don't think a second server is needed. :P Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 6 Apr 08 Posts: 2018 Credit: 100,142,856 RAC: 0 |
I expect the top crunchers amassed most of their credits before the ATIs were available for optimized crunching. ATI GPU crunching is relatively quite recent and I expect people want a stable period of crunching with the ATIs before saying how they are doing with them. |
Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0 |
Agreed, not all CC's will dump the whole project directory on a reset. However, all will dump the stock app on one. However, that wasn't the real point. The new GPU apps are bandwidth and work sponges from the Projects POV. Therefore, 'casual' resetting forces those hosts to pull all the input files and a whole new set of work from empty, with the attendant DB and and bandwidth load that goes along with that. Given that the project is having trouble keeping up as it is, anything that increase that even more only makes things worse for the big picture. Also, there is no guarantee that the scheduler queue hasn't been sucked dry just as the host gets reset, so the whole process would have been done for no net gain in that case. Alinator |
Send message Joined: 17 Jan 09 Posts: 98 Credit: 72,182,367 RAC: 0 |
I know have work on all computers and including the GPU where I was getting practically nothing. |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
I know have work on all computers and including the GPU where I was getting practically nothing. Server-side, it looks like the change helped work availability quite a bit. |
Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0 |
I know have work on all computers and including the GPU where I was getting practically nothing. LOL... Well, we had our three steps backwards over the last 36 hours or so. So I guess that would be about 2 1/2 forward now? ;-) Alinator |
Send message Joined: 9 Sep 08 Posts: 96 Credit: 336,443,946 RAC: 0 |
I know have work on all computers and including the GPU where I was getting practically nothing. The only problem- and it's a big one- is when you have 'got 0 new tasks' multiple times in a row, which does still happen, and run dry and the reconnect time goes to 2 or 3 hours and you have nothing to crunch until then... any way to change/fix(?) that? |
Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0 |
I know have work on all computers and including the GPU where I was getting practically nothing. It depends on what mean by 'fix'. If you are talking about MW being able to feed work to the scheduler faster, then no, not until Monday at the earliest. If you're talking about not running dry, then you have two choices: 1.) Sit at the console all the time and pound on the update button when you are out. 2.) Run a backup project and just ride out these difficulties until a satisfactory backend fix can be implemented. Alinator |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
I think he was talking about the 3 hour delay. Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 20 Mar 08 Posts: 108 Credit: 2,607,924,860 RAC: 0 |
So far, I've seen *none* of the 0 responses I spoke too soon. Quite a few 0 responses hidden in the logs, and running dry on one occasion. But still a significant improvement over both the previous per-core quotas tried, as far as my crunchers are concerned. |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
So far, I've seen *none* of the 0 responses Should be even better now that the workunits should take around twice as long to crunch. |
Send message Joined: 6 Apr 08 Posts: 2018 Credit: 100,142,856 RAC: 0 |
|
Send message Joined: 9 Sep 08 Posts: 96 Credit: 336,443,946 RAC: 0 |
'If you are talking about MW being able to feed work to the scheduler faster...' I'm not. 'I think he was talking about the 3 hour delay.' I was. |
Send message Joined: 15 Jan 09 Posts: 169 Credit: 6,734,481 RAC: 0 |
Good to see that the idea of increasing the size of a work unit was eventually taken on board. Making them smaller always seemed crazy when they are mostly completed so fast :) |
Send message Joined: 27 Aug 07 Posts: 915 Credit: 1,503,319 RAC: 0 |
Should be even better now that the workunits should take around twice as long to crunch. Ooh 2 seconds then. me@rescam.org |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
Should be even better now that the workunits should take around twice as long to crunch. Hey now, some of the GPU apps are taking a whole 3 seconds. |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
Should be even better now that the workunits should take around twice as long to crunch. Need a faster app now. Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
©2024 Astroinformatics Group