| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: The Clean Energy Project - Phase 2 Forum
Thread: Poor crunching on this project |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 156
|
|
| Author |
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges: 
|
Actually had 2 problems with this project. I have an 8 core pc running ubuntu. I switched to this project and noticed my crunching time wasn't a full 24 hrs crunching per core. Not sure if it was stalling for some reason. Another problem was uploading. I have a DSL connection. Since the data was pretty large, I restricted uploads to 11pm to 6 am. My completed wu's would not get uploaded during this time frame due to server problems on the receiving end. I watched a few times and saw the uploads stall, then start again. The message log stated my connection was fine but problems on the other end. Ended up aborting this project and went back to clean water. Anyone else have problems like this? Would like to get at least a gold badge on this project.
----------------------------------------
Crunching for humanity since 2007!
 |
||
|
|
gb009761
Master Cruncher Scotland Joined: Apr 6, 2005 Post Count: 2955 Status: Offline Project Badges:
|
Looks like you've already got a Gold badge for this...
----------------------------------------As to problems with run time/uploading - I'm awaiting the Windows version, thus, I'll leave that for others to comment on...  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I have the exact same setup as you and CPU efficiency only reaches about 95% with this project. I believe it has something to do with the heavy disk activity involved with these work units. The CPU has to wait for data to be written/retrieved from the relatively slow hard drive so some efficiency is lost. You can minimize this by running a mix of projects so that no more than 3-4 Clean Energy work units are running at the same time. I specifically installed Ubuntu just to crunch this project (I'm a Windows user normally) so I just live with the 5% loss.
My machine only connects from 1am-6am local time but I have not noticed any upload problems in the message logs. I only have 65KB/s upload bandwidth so it takes the machine over 2 hours to send it's payload, but there's never been a problem that I've noticed. Everything is always uploaded by morning. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Actually had 2 problems with this project. I have an 8 core pc running ubuntu. I switched to this project and noticed my crunching time wasn't a full 24 hrs crunching per core. Not sure if it was stalling for some reason. Another problem was uploading. I have a DSL connection. Since the data was pretty large, I restricted uploads to 11pm to 6 am. My completed wu's would not get uploaded during this time frame due to server problems on the receiving end. I watched a few times and saw the uploads stall, then start again. The message log stated my connection was fine but problems on the other end. Ended up aborting this project and went back to clean water. Anyone else have problems like this? Would like to get at least a gold badge on this project. Hi Bearcat, When did you make that experience? The uploading restarting due time out's I've not seen for longer and mix CEP2 in with other WCG sciences on a continuous basis. The _4 result part is presently about 23MB and they upload here uninterrupted in about 6 minutes when one at a time (the CEP2), with a BOINC pref upload limit to 84KB. Full 24hrs you wont get with any project in mean, not even the lightest simply because your system itself uses time. Very best I ever see is 99.8% for Clean Water/HCC/HCMD2 and others. For CEP2, if I leave the host alone it's about 30 minutes per job that go towards i/o. Big model and BOINC simply not counting that time. It's a concern the techs try to minimize before launching on Windows and then suspect that any improvement will be ported back to the Linux version.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Oct 2, 2010 6:06:42 AM] |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
Actually had 2 problems with this project. I have an 8 core pc running ubuntu. I switched to this project and noticed my crunching time wasn't a full 24 hrs crunching per core. Not sure if it was stalling for some reason. Another problem was uploading. I have a DSL connection. Since the data was pretty large, I restricted uploads to 11pm to 6 am. My completed wu's would not get uploaded during this time frame due to server problems on the receiving end. I watched a few times and saw the uploads stall, then start again. The message log stated my connection was fine but problems on the other end. Ended up aborting this project and went back to clean water. Anyone else have problems like this? Would like to get at least a gold badge on this project. If you limit your usable upload bandwidth to 80% for this project your upload problems should go away. Multiple completed WUs trying to upload at the same time will cause time outs unless you have a really fast connection with at least a 1 meg upload speed. Even then I would still recommend an 80% throttle. Works for me.
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.
  |
||
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges:
|
My total run time on this project should be 192 hrs when running none stop with my 8 cores but only hover around 100 to 125 hrs per day. I can see a little drop when it's writing to disk but not that much. Will have to check my disk to see if it's spinning down. I wish boincwould only upload 1 at a time because I did see allot of time outs. Didn't think I had enough to get gold but glad I did. Thanks for the help.
----------------------------------------
Crunching for humanity since 2007!
|
||
|
|
RaymondFO
Veteran Cruncher USA Joined: Nov 30, 2004 Post Count: 561 Status: Offline Project Badges: 
|
Two of my computers have recently experienced poor crunching results with this project, and nothing is helping. Running a mix of other projects, upgrading the BOINC client and OS and nothing is working. Here is an example:
 Before it ran with C4CW WU's, the CPU versus wall clock time differential was larger than other projects (20-40 min.) but still acceptable. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I have just one PC (single core) doing CEP2 and it's on 24x7. I started crunching around 30-Jun or 01-Jul and I just got my gold badge (90 days) so in my case it seems that elapsed and cpu time are almost identical.
|
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Two of my computers have recently experienced poor crunching results with this project, and nothing is helping. Running a mix of other projects, upgrading the BOINC client and OS and nothing is working. Here is an example: http://i1047.photobucket.com/albums/b479/RaymondFO/Screenshot.jpg Before it ran with C4CW WU's, the CPU versus wall clock time differential was larger than other projects (20-40 min.) but still acceptable. I was digging into process affinity for the first time for Ubuntu and wondered if something is funky with that in Linux with Hyperthreading. Since CEP2 has multiple threads open, do several BOINC sciences land on the same core?
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
RaymondFO
Veteran Cruncher USA Joined: Nov 30, 2004 Post Count: 561 Status: Offline Project Badges:
|
Sekerob,
I am not sure if several BOINC sciences land on the same core does or does not occur. The computer was exclusively running CEP2 for over two months, with the exception when the computer was blessed to receive a DDDT2 WU. After DDDT2 supply dried up, I started to run C4CW to replace the occasional DDDT2 WU when this started to happen on all CEP2 WU's. Another computer that does not have hyperthreading capabilities (Intel Q9400 chip) also experienced the same issue. Wall clock would be 6-7 hours, processor time usually around 2.75 hours. Any ideas or request for additional information is welcome. |
||
|
|
|