1)
Message boards :
Number crunching :
Download Problems
(Message 628)
Posted 25 Feb 2006 by Ingleside Post: BTW: I downloaded 5.3.20 from BOINC home page. Is there any info about the version history ? "version details" on download-page atleast contains partial info... |
2)
Message boards :
Feedback :
Credit scores
(Message 529)
Posted 23 Feb 2006 by Ingleside Post: As a developer Ingleside, should know that. Hmm, looks like become too hung-up with BOINC alpha, but when again their instructions clearly states "If you have experienced problems with BOINC that are not exercised by any of these tests, please post to the email list." Anyway, RALPH@home homepage states: February 15, 2006 Hmm, looks like the translation got screwed-up (again), since my interpretation of "improve Rosetta@home" meant all of Rosetta@home and not just to check if the latest bug-fixes worked. :oops: In any case, this thread has become a waste of time for everyone, so just nuke it. |
3)
Message boards :
Number crunching :
Download Problems
(Message 490)
Posted 22 Feb 2006 by Ingleside Post: I'm running WinXP SP2 and CC 5.2.13 (optimized version), work kept in memory. Are you behind a proxy? If so, the most likely reason is the download-bug in BOINC v5.2.x that is fixed in v5.3.15... |
4)
Message boards :
Number crunching :
Resource share: RALPH instead of Rosetta@home?
(Message 488)
Posted 22 Feb 2006 by Ingleside Post:
BOINC v5.3.18 added support for using global_prefs_override.xml, this will override any of the global preferences you specifies on the computer you've added this file. But, the BOINC Manager haven't yet got support to re-read this file, so only works by re-starting BOINC. So, if example normally leaves application in memory but wants to override on the RALPH-computer, just make yourself a global_prefs_override.xml with: <global_preferences> <leave_apps_in_memory>0</leave_apps_in_memory> </global_preferences> |
5)
Message boards :
Feedback :
Credit scores
(Message 482)
Posted 22 Feb 2006 by Ingleside Post: RALPH is a ALPA testing project. Credits are a total non-issue for RALPH. This question would better be asked in the Rosetta Forums, as this project will not be addressing how credits are granted in BOINC. RALPH is the alpha-testing project for Rosetta@home, and its point will therefore be to test any improvements/changes to Rosetta@home before it's released on the "public" project, to make sure the Rosetta@home application works and doesn't crap-out most of the time. In Rosetta@home there's been numerous threads discussing the problems with the current crediting, and possible ways to improve crediting. Let's take a quick look on the methods BOINC has to improve crediting: 1; Using redundancy, there granted credit is either lowest claimed if 2 validated, or remove highest and lowest and average the rest if 3 or more. Since Rosetta@home has decided they don't need redundancy to make sure the science is correct, using redundancy will not be a good method since this gives lower throughput. 2; Using fixed credit for each wu-type, like CPDN and Folding@home is doing. But, in Rosetta@home different wu from same wu-type can still give huge differences, so isn't really usable. 3; Using boinc_ops_per_cpu_second, application-specific benchmark. Any project wanting to use this must add a customized benchmark to their application. 4; Using boinc_ops_cumulative, application "counts flops". Seti_Enhanced is using this method, but just like for boinc_ops_per_cpu_second it means any project wanting to use this must add code to their application to "count flops". So, the BOINC-developers has supplied the neccessary support in BOINC-client and server-backend to give better granted credit, that an individual project isn't taking advantage of this support isn't really anything the BOINC-developers can do much with. This basically means, the Rosetta@home-developers has 3 options, starting to use redundancy that they've already said they won't do, don't do anything but at the same time risk alienating many users that gets fed-up with the broken credit-system and therefore hurt the science, or add boinc_ops_cumulative that a couple of their developers already have hinted to atleast looking into... But, adding boinc_ops_cumulative means code-changes to the application, meaning any likely change to Rosetta@home's crediting should be alpha-tested by RALPH before it's released, and is therefore on-topic in this alpha-project. |
6)
Message boards :
RALPH@home bug list :
Same team on 20th and 21th rank in teams list
(Message 212)
Posted 18 Feb 2006 by Ingleside Post: Hmm, looks like there is some mess in teams list. As I see it now the L'Alliance Francophone team (teamid=24) is also on 20 and 21 rank. Adding to what Mchl has already mentioned, to lower the load on BOINC-database, team-pages and so on is generated when needed and cached on web-server for some time. This means different pages isn't generated at the same time, and this gives instances there a team, computer or user is either showing-up on 2 pages, or ocassionally on none of the pages at all... |
©2024 University of Washington
http://www.bakerlab.org