| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Community
Forum: Chat Room
Thread: Okay, I confez, I'm not a Linux Adapt. |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 581
|
|
| Author |
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges: 
|
Any links providing information to prove this?
----------------------------------------
Crunching for humanity since 2007!
 |
||
|
|
Dark Angel
Veteran Cruncher Australia Joined: Nov 11, 2005 Post Count: 728 Status: Offline Project Badges: 
|
Just look at the facts for yourself.
----------------------------------------Linux box finishes given work units faster than Windows. Units are the same size across platforms. Windows claims higher efficiency but still takes longer to finish work units on identical hardware. ergo - Windows efficiency metric is not the same as Linux metric. Also, it is in MS's interests for Windows to report the highest feasible efficiency (or any other metric) Any admission of lower performance might give credibility to alternative systems. Lower performance has already been established but Windows claims higher efficiency, ergo Windows is either calculating differently to give the highest figure it can or it's outright lying (padding the figure).  Currently being moderated under false pretences |
||
|
|
KWSN - A Shrubbery
Master Cruncher Joined: Jan 8, 2006 Post Count: 1585 Status: Offline |
I think you will find that most people on these forums who discuss efficiency are pulling the numbers from BOINCTasks. As this compares elapsed time vs. cpu time as reported by the client, I don't see how Windows can "massage" the numbers.
---------------------------------------- Distributed computing volunteer since September 27, 2000 |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I must say that I have seen many people go back on this for quite some time. I myself think that Linux is slightly faster, although less efficient, (CEP2).
TO ME Windows is a "heavier" operating system, and therefore MAY use more CPU cycles in background. The more "stuff" that's going on in the background in my opinion should slow the WU's down (during CPU usage), thus allowing for a higher efficiency since the HDD or SDD can keep up with it. Nothing besides one rig I've flip-flopped with to back this up though. Just thoughts. |
||
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges:
|
Both of my rigs use the same motherboard, super micro X8DTI-F. My main rig uses Win7 with 2 X5670 xeon processors and 4gb sticks ram totally 24gb. The ubuntu rig uses 1 X5650 (other socket empty) and 2gb sticks for 12gb ram. Besides processor clock speeds, should be to much issues with difference with CPU vs actual time. Would think Win7 would have more differences with all the extra crap running in the background vs Linux, but doesn't seem to be this case. Might try running boinc on a separate hard drive (mechanical) and leave ubuntu on the ssd to see if any difference. Tried just about everything else so one more try to keep Linux running before I switch.
----------------------------------------
Crunching for humanity since 2007!
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Bearcat,
Setting efficiency aside, which one completes WU faster? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
So back on my inquiry, anyone having any experience to share about the "ionice" option?
TIA --//-- |
||
|
|
marvey11
Advanced Cruncher Germany Joined: Apr 2, 2011 Post Count: 89 Status: Offline Project Badges: 
|
SekeRob,
----------------------------------------I don't know ionice (yet?) but I had a look at the link which contained the man page. AFAICS you need to specify the process ID as the argument to the -p option and not the executable's name. So you might try using the actual PID or omit "-p". HTH, and please report the outcome -- I'm also interested in getting CEP2 a bit more efficient. -- MW  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
As I eluded, with PID it works and scripts could be run to discover the PID and then feed it into the command, But, the function does allow putting the program name in place, then put this in the updatedb so they'll always run at the desired disk io priority. (CEP2 and VINA jobs constantly change PID, me preferring the KISS). After doing the Re-RTFM some more I realized that -p has to be replaced by -n plus a priority from 0-7 and the program, so this works:
sudo ionice -c 3 -n 3 wcgrid_cep2_6.40_i686-pc-linux-gnu and this too sudo ionice -c 3 -n 5 wcgrid_cep2_qchem_6.40_i686-pc-linux-gnu But, what do those having experience with this have to say, is my eagerness to learn? --//-- |
||
|
|
marvey11
Advanced Cruncher Germany Joined: Apr 2, 2011 Post Count: 89 Status: Offline Project Badges:
|
Welll, there's always the pidof command.
----------------------------------------marvey@helium:~$ pidof wcg_hpf2_rosetta_6.40_i686-pc-linux-gnu But if certain WCG projects keep switching through sub-processes, this won't help you, either... |
||
|
|
|