Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
![]() |
World Community Grid Forums
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 35
|
![]() |
Author |
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
While I only have a few PV's and Valids on my i7-2600K Ubuntu 11.04-x64 system, it is outperforming a similar Windows-x64 system by some way:
2.6hours on average for Ubuntu v's over 4.6h on average for Windows. That would make Linux 77% faster - a bit more significant than I would have expected. Anyone else seeing similar relative performances? Are the task lengths varying from day to day? |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
While I only have a few PV's and Valids on my i7-2600K Ubuntu 11.04-x64 system, it is outperforming a similar Windows-x64 system by some way: 2.6hours on average for Ubuntu v's over 4.6h on average for Windows. That would make Linux 77% faster - a bit more significant than I would have expected. Anyone else seeing similar relative performances? Are the task lengths varying from day to day? The variability of the tasks that I've seen, from 1.97 to almost 4.98 hours, median 4.2 hours under Linux 64 bits makes me think that your conclusion is rather quick. Run a hundred on each platform and you might have a good number to talk with. --//-- |
||
|
marvey11
Advanced Cruncher Germany Joined: Apr 2, 2011 Post Count: 89 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I have no comparison results for you since I'm running Linux machines only - - one of them 24/7 and the other only weekdays from 9 to 5
----------------------------------------![]() What's perhaps more important for you to compare is how fast steps (jobs?, tasks?) of the same size are crunched on the different systems. What makes me think so are the vast differences in CPU times I can sometimes see in stderr.txt for crunched steps within the same result (see example below). Therefore I think that a result's final CPU time depends in large parts on the steps' sizes and their distribution within a result. So it might simply be a coincidence that your Ubuntu cruncher has snagged a few results with comparatively smaller step sizes. BTW, the number of steps in a result always seems to be a multiple of 4, and there are always groups of 4 consecutive steps of the same size (sometimes even 8 or 12, but I simply consider those as being several groups of 4 steps each). And the second number of the job's "size" tuple seems to have a larger impact on CPU times than the first. Example from DSFL_ 00000001_ 0000000_ 0701_ 3: [14:14:34] Number of tasks = 28 ![]() |
||
|
kateiacy
Veteran Cruncher USA Joined: Jan 23, 2010 Post Count: 1027 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I'll bite. I have one dual-boot machine (all my others are Linux only).
----------------------------------------I'll run exclusively DSFL on the Windows side for a few days, then switch over and do nothing but DSFL on Linux. Will report back. ![]() |
||
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
If that's the case, I'll be switching my win7 rig back to ubuntu. Any errors on ubuntu?
----------------------------------------
Crunching for humanity since 2007!
![]() |
||
|
kateiacy
Veteran Cruncher USA Joined: Jan 23, 2010 Post Count: 1027 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
DSFL has been running fine (no errors or invalids) for me on 4 machines under Ubuntu (3 on 10.04.3 and one 11.04). The betas for it did as well, once we got past the very first batch that had the problems with running wild and overheating on Linux and with AMD/Intel mismatches.
----------------------------------------![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
both of my intels are still running warm. reduced the cpu usage down to 80% of i7 and 60% on i5 still hitting 81c and 73c respectively.
|
||
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I have seen no DSFL errors so far from either Linux or Windows systems.
My average Windows times are down slightly to 4.4h and my average Linux times went up to 3.25h, making Linux more like 35% faster, if these are reasonably accurate representations (just 50tasks) - probably need a couple more days to be sure, and I would expect it to fall slightly more to perhaps 25 to 30%. Anyway, this is a lot closer to what I would have expected. I did not notice an increase in task length for Windows, so I expect tasks were not generally increased in size, unless they were just increased for Linux (and I doubt that). So the tasks may just be somewhat random in size and I initially got oddly low task lengths on my Ubuntu system. I suppose my Ubuntu i7-2600K may have run dry-ish at some stage, exasperating the delta; running just a couple of tasks would have meant a higher turbo frequency, increasing individual task performance by around 15%. [OT] txyankee, 81 deg C at 80% is too high; unless you are using an Intel heatsink! Check it's pinned/clipped to the board correctly and there is no heavy build up of dust. Sometimes just moving a computer out of a corner can reduce temps by around 5deg. |
||
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
both of my intels are still running warm. reduced the cpu usage down to 80% of i7 and 60% on i5 still hitting 81c and 73c respectively. Have never had any project raise my temps much before so yours is a bit weird. I suspect you have dust build up, or you need to reapply thermal paste to your heat sinks. Your temps are to high. Are these desktops or laptops?
Crunching for humanity since 2007!
![]() |
||
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
both of my intels are still running warm. reduced the cpu usage down to 80% of i7 and 60% on i5 still hitting 81c and 73c respectively. If you're running the stock HSF that is probably your problem. They are pretty much useless. If it's not that check for dust bunnies and/or air flow. What are your ambient temps?
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
![]() ![]() |
||
|
|
![]() |