| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Beta Testing
Forum: Beta Test Support Forum
Thread: Beta CEP2 Mac and Linux test version 6.37 |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 67
|
|
| Author |
|
|
Crystal Pellet
Veteran Cruncher Joined: May 21, 2008 Post Count: 1406 Status: Offline Project Badges: 
|
Using Linux 2.6.32-21-generic, I did not notice a low efficiency:
BETA_E200530_022_A.25.C20H10OS3Se.21.0.set1d06_0 09:16:36 (09:07:52) BETA_E200530_055_A.25.C22H16N2O.1.4.set1d06_0 08:34:36 (08:24:28) BETA_E200530_015_A.25.C22H16N2O.10.3.set1d06_1 08:08:05 (07:58:15) BETA_E200530_023_A.25.C22H16N2O.10.4.set1d06_0 04:40:34 (04:31:23) Is it possible that testers who are noticing low efficiency have "Leave application in memory while suspended" ticked off and that tasks were suspended for whatever reason, so that after resuming that Job has to restart from the last checkpoint and that could be 2 hours loss of CPU-time? |
||
|
|
Anarki
Cruncher Joined: Jan 25, 2007 Post Count: 34 Status: Offline Project Badges: 
|
Is it possible that testers who are noticing low efficiency have "Leave application in memory while suspended" ticked off and that tasks were suspended for whatever reason, so that after resuming that Job has to restart from the last checkpoint and that could be 2 hours loss of CPU-time? I have never noticed this setting until you just mentioned it, should the box be ticked or unticked? What does this setting mean/do? |
||
|
|
Crystal Pellet
Veteran Cruncher Joined: May 21, 2008 Post Count: 1406 Status: Offline Project Badges:
|
With tasks with checkpoints so rarely as CEP2 it is advisable to tick that box.
----------------------------------------If the task is suspended (running Boinc-benchmark?, manual?, other project running High Priority?), the state of the process stays in memory (or swapped to disk), so that after resuming, the task can start where it was without loss of CPU-time. It of course don't work when rebooting or restarting the BOINC-client. On laptops it also should work when the laptop is going into Hibernate by closing the lid. [Edit 1 times, last edit by Crystal Pellet at Nov 17, 2010 11:54:45 AM] |
||
|
|
Anarki
Cruncher Joined: Jan 25, 2007 Post Count: 34 Status: Offline Project Badges:
|
Thanks for the explaination, it was ticked on my client so I'm guessing the lost CPU time is down to something else.
|
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: 
|
For the users seeing the very low efficiency is this only with running multiple workunits at the same time? Is anyone seeing this low efficiency while only running one CEP2 workunit?
Thanks, armstrdj |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Only shows on the 16th Job end and probably linked to the effort to home in on the lost CPU time. Opposed to Windows that has 2 processes per CEP2 job, Linux has 4 with changing PIDs on each job. I see 6 processes. 3x "wcgrid_beta11_6.37_i686-pc-linux" and 3x "../../projects/www.worldcommunitygrid.org/wcgrid_beta11_qchem_6.37_i686-pc-linux-gnu". The PID reported is of the first instance of "wcgrid_beta11_6.37_i686-pc-linux", which is parent (or grandparent, etc) of all the others. Meanwhile, I've had plenty of betas, all but one of which have validated successfully. The other was the usual "RC = 0x4" problem that afflicts P3-family machines that make it to job #4. |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges:
|
Starting to pick up a few repair WUs that errored out on other machines. They seem to be doing ok on my machines so far. *fingers croassed*
----------------------------------------
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.
  |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
6 indeed... is that a change from the production versions? The structure in System Monitor looks like 4 and 2.
----------------------------------------Well, surely you read the reply by cleanenery... long as 1 job is completed it's a useful result, but "P3-family machines that make it to job #4." read ominous. [ot]Dont they frequently make it to job 4 and then the question being where a wingman could be getting all the way to 16? My P4 has long been retired... it still boots at times to run some legacy XP stuff that wont work even in 32 bit emulation under W7. The watts "energy" to output ratio is just atrocious compared to any newer CPU and I could only justify if the machine is needed to be on for actual "user" purpose, not a second longer. None of the WCG tasks manage to run on that in 2 hours, so it doesn't.[/ot]
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
codes
Advanced Cruncher Joined: Oct 20, 2009 Post Count: 142 Status: Offline |
Is anyone seeing this low efficiency while only running one CEP2 workunit? I wouldn't call it low efficiency on my device, the BETA has the same typical CPU and elapsed times (12 hours) as the production WU on my single core Intel Celeron 1.8 GHz running Linux. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Well, surely you read the reply by cleanenery... long as 1 job is completed it's a useful result, but "P3-family machines that make it to job #4." read ominous. [ot]Dont they frequently make it to job 4 and then the question being where a wingman could be getting all the way to 16? My P4 has long been retired... it still boots at times to run some legacy XP stuff that wont work even in 32 bit emulation under W7. The watts "energy" to output ratio is just atrocious compared to any newer CPU and I could only justify if the machine is needed to be on for actual "user" purpose ...[/ot] It doesn't matter if the result is useful or not, because a P3 WU is marked invalid if it gets to job 4. If it only completes 0/1/2/3, it will be marked valid. And, yes, a P3/1.4 frequently makes it to job 4. As for the ot part: I have a P3/1.4 file server that is always running. Sending files onto ethernet takes very little CPU power, so there's really no point upgrading it. It's quiet and cool and almost all faster CPUs would simply use more power (esp. P4) to do the same job. |
||
|
|
|