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: 54
|
![]() |
Author |
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
nano, you are getting about 16 pts per hr under linux. Are you saying you getting the same under windows? I'm sure HT does skew the numbers. You have twice the cores running vs. a quad but you don't get twice the points or WUs completed. I do get more points per hour running Windows but I don't think it's more than 10 or so. It varies with the projects run. HCC seems to be the lowest because it is the fastest.
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.
![]() ![]() |
||
|
anhhai
Veteran Cruncher Joined: Mar 22, 2005 Post Count: 839 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I have HT turned on too. Here is an easy experiment, run a c4cw WU on your linux box (or any of the zero redundancy projects).
----------------------------------------edit : c4cw is the only one I know for a fact is faster under linux (3.2 hrs vs 4.7 under my machine) ![]() [Edit 1 times, last edit by anhhai at Nov 15, 2010 2:04:56 PM] |
||
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I will run some C4CW on the 920 but I won't have any Windows numbers to make a comparison.
----------------------------------------
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.
![]() ![]() |
||
|
anhhai
Veteran Cruncher Joined: Mar 22, 2005 Post Count: 839 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
the zero redunancy once shouldn't matter in terms of pts per hr. you may get less pts per WU, but that is because you finish them faster. This is why I prefer to run mostly only HCC and c4cw on linux. They are faster then windows.
----------------------------------------The pts scheme to me should tell me how efficient my system is at running a certain project. For example, on 1 of my systems, I would regularly claim 2X the granted pts when I run HPF2. However, for the same system, I would always claim too few credits on all of the zero redundancy projects and not off by much on the other projects. So I stop sending HPF2 to that system. ![]() |
||
|
krakatuk
Advanced Cruncher Germany Joined: Oct 3, 2008 Post Count: 141 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Anhhai, did you try CEP2 in Dotsch from a USB stick?
----------------------------------------Would be interesting to know if it runs there and how is the performance. ![]() |
||
|
anhhai
Veteran Cruncher Joined: Mar 22, 2005 Post Count: 839 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I run it from a HD. I am getting between 97-98.5% efficiency. However, I limit it to just 4 (out of 8) WU at a time. Also, some times I don't run CEP2 at all on the faster systems. For my windows system (16 cores), I limit it to just 2 at a time. I am pretty sure it can handle more, but I actually do work/games on these system, so I don't want to stress them too bad.
----------------------------------------However, overall I am not really focusing on CEP2, too much effort to manage. ![]() |
||
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
"Which Linux is best?", well kernel 2.6.38 is supposedly ripping something from the BFS scheduler and getting an integration which should make things super smooth. Read http://www.phoronix.com/scan.php?page=article...inux_2637_video&num=1
----------------------------------------Who knows will that also cut the waiting on Disk IO states of CEP2 jobs, if the research has not finished by then ;? As yet, can't remember which Kernel is going to be used on Nasty Narwal (seemingly release postponed), .36 or .37 ?
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Anhhai, did you try CEP2 in Dotsch from a USB stick? Would be interesting to know if it runs there and how is the performance. I tried it when CEP2 first came out running it on quads and it killed the USB drives. Well not really killed them but corrupted them so bad I had to format and reinstall. Too many read/writes for them too handle on a multicore. I found out later that if you use the restriction in the device manager to limit CEP2 to 1 per core I have had them run on USB with no issues. They did OK on the betas too but I had to make sure only 1 ran at a time because the betas seem to ignore the 1 per core override.
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 |
On the theme of best [performant] Linux, an extensive test of the last 26 Kernels shows that Disk IO is vastly better from release 2.6.30. Given that I'm frequently seeing lower kernel versions in CEP2 quorums. Just a suggestion :D
----------------------------------------![]() Test source: http://www.phoronix.com/scan.php?page=article...linux_2612_2637&num=7 Now, I have a USB 3.0 PCIe card with 1 free slot and the shops are waiting on delivery of the 32GB LaCie USB 3.0 stick which has a read speed of 260Mb... which will fit the budget. Kingston is coming too with a USB 3.0 SSD but these are hefty in price with 195Mb W/R. crunching on --//--
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
kateiacy
Veteran Cruncher USA Joined: Jan 23, 2010 Post Count: 1027 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
On the theme of best [performant] Linux, an extensive test of the last 26 Kernels shows that Disk IO is vastly better from release 2.6.30. Given that I'm frequently seeing lower kernel versions in CEP2 quorums. Just a suggestion :D Thanks for the useful information! ![]() |
||
|
|
![]() |