| 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: 96
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi SekeRob,
thanks for reminding us ;). Cheers Your Harvard CEP team |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Last night did go again for the all stops out ''all cores'' concurrent trick on the 8 threaded I7 and started them with intervals of 3 minutes to get them passed the 1st checkpoint and this morning all 8 sat ready in the transfer tab, for the scheduler to open the network at 10:3. Per BOINCtasks they ran at 98.2 to 98.9% efficiency in hands off.
Meantime, the crunching chart http://bit.ly/WCGCE1 has been showing since yesterday night the nice round number of 10 million completed results... Way on to the next 10M, and sooner please. Crunch On --//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Last night did go again for the all stops out ''all cores'' concurrent trick on the 8 threaded I7 and started them with intervals of 3 minutes to get them passed the 1st checkpoint and this morning all 8 sat ready in the transfer tab, for the scheduler to open the network at 10:3. Per BOINCtasks they ran at 98.2 to 98.9% efficiency in hands off. Just out of curiosity, did you watch your core temps running all cores/threads? I usually run I7s at (cores/threads - 1) because "I'm scared of heat" (air-cooled). On the bright side, get about 99.5% or higher efficiency (RAID drives) even with running a Collatz GPU-only simultaneously. Crunch On --//-- And...will BOINCTasks do the staggered task startup for you? If yes, I'll RTFM. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It makes not one iota of difference running 7 or 8 threads on the total temps. If you get 99.5% with 7 and I get a mean of 98.5% with 8, no idle threads, you can do the sums how much more work get's done with 1% efficiency loss on 8 concurrent... that's about 13% more done with maybe 1C difference in op-temps. Once the staggered start is complete, continuing to run more pretty much makes it off-chance that any more will start simultaneous during that critical setup phase to get to the first checkpoint. Explicitly, this is with I7 and W7-64 on a single disk.
No, there is no miracle feature in BOINCTasks that lets you suspend tasks in fixed time intervals to let the CEP2 start... it's hands on, extremely easy in BT to monitor the checkpoint taking to start a series in steps, and pick a time to boot. --//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Last night repeated the all 8 threads concurrent exersize, but started early whilst working on some charting stuff. This morning the 8 were listed with 97.2 to 98.5% efficiency. Not bad at all, but occasinally did spot a little lag, probably during the disk IO parts [there's only a case internal HD light that cant be seen]. Added another 62 hours to the tally, making CEP2 me second largest contributed research...
--//-- P.S., yesterday all volunteers added overall 24.4 years to the total [including the 5 hours or so validation delay of Monday]. We need 25 more to beat the January month record. |
||
|
|
mecole
Advanced Cruncher Joined: Jun 18, 2011 Post Count: 55 Status: Offline |
Thanks for the continual updates on efficiency ... I have been micro managing loading the cache so I get 1 CEP2, 2 SN2S, 1 CEP2, 2 SN2S, etc until I build up a couple of days worth for. This keeps my concurrent CEP2 at 3-4 but seeing your efficiency study show that 8 threads is not really bad at all I am thinking I might just let my i7-920 on Win7 64 run wild <party>
----------------------------------------I'm a bit lazy ... any efficiency studies on 8 threads WinXP 32? ![]() |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes, thanks for the updates from us as well!
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Considering how hard our devices have to 'sweat' on these, Ten-thousand CPU years is major!
http://bit.ly/WCGCE1 Crunch On --//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Absolutely, Sekerob! We'll also use the opportunity to promote your chart in one of our next logposts ;).
Cheers Your Harvard CEP team |
||
|
|
Platoon
Advanced Cruncher Russia Joined: Jun 28, 2006 Post Count: 62 Status: Offline Project Badges:
|
Just bumping up
----------------------------------------
" forever forge ahead and keep the dream in sight!"
![]() |
||
|
|
|