| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: WCG cross-project performance benchmark |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 3
|
|
| Author |
|
|
KerSamson
Master Cruncher Switzerland Joined: Jan 29, 2007 Post Count: 1684 Status: Offline Project Badges: 
|
Based on BoincTasks report, I took the time to generate a brief performance analysis between ZIKA, MCM, and SCC vs operating systems and CPU architectures, see below picture.
---------------------------------------- (high resolution picture ) Since the machines are running 24/7 and devoted to WCG, I am really surprised to notice such significant performance differences (about 30% difference). Should the boinc CPU benchmark be revised / improved? We know that the Windows memory management is poorly efficient. It explains the less good performance of Windows-based hosts comparing to Linux-based hosts. Nevertheless, why is the cross-project performance of Windows-based hosts relatively stable when we have to notice significant differences by Linux-based hosts? My remark is not related to any boinc credit competition - since I do not participate to any of such challenges - but only related to the real project performances. Cheers, Yves --- NB: Between 2019-07-09 and 2019-07-13, kermorvan did experience some troubles and the host was not 100% available for WCG. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yves,
SCC and Zika are both considerably more efficient under Linux than Windoze. It has been discussed endlessly elsewhere in the fora. |
||
|
|
KerSamson
Master Cruncher Switzerland Joined: Jan 29, 2007 Post Count: 1684 Status: Offline Project Badges:
|
I'm aware about the situation regarding SCC and Zika on Linux.
----------------------------------------My questions are more focused on / related to MCM on Linux. |
||
|
|
|