Posts by Spare_Cycles
|
1)
Message boards :
Current tests :
New crediting system
(Message 2010)
Posted 12 Aug 2006 by Spare_Cycles Post: I am eager to see the fair credit system on Rosetta, but first we need to make sure that the current server problems on RALPH are not somehow being caused by a bug in the new credit system. It's not possible to have two official stat systems. The stats that are exported to all the third party stat sites are the official stats. To attract crunchers who want to compete with fair stats, these exported stats must be the fair stats. |
|
2)
Message boards :
Current tests :
New crediting system
(Message 1942)
Posted 9 Aug 2006 by Spare_Cycles Post: If inter-project parity is one of the objectives, I'm not seeing it yet. Inter-project parity is not an objective on RALPH and thus it may never happen on RALPH. A credit of 2 per model is being given no matter how slow or fast the WU in order to test the awarding of credit. As far as I can tell, the test is working fine and credit is being awarded as intended. |
|
3)
Message boards :
Cafe RALPH :
SOMEone PLEASE vote SOMEONE for user of the day!!
(Message 1448)
Posted 1 May 2006 by Spare_Cycles Post: just seems stuck on the man from manchester. I think you're right. Whatever bit of software that is supposed to update the "user of the day" doesn't seem to be running. |
|
4)
Message boards :
Cafe RALPH :
SOMEone PLEASE vote SOMEONE for user of the day!!
(Message 1384)
Posted 26 Apr 2006 by Spare_Cycles Post: Ok, I just voted for someone named "feet1st" for user of the day. So perhaps he will become "user of the day" for the next month or so. |
|
5)
Message boards :
Number crunching :
Team stats not adding up...
(Message 871)
Posted 13 Mar 2006 by Spare_Cycles Post: I believe the team only gets credit while you are crunching for them. Credit that you had before you joined the team isn't added to the team total. |
|
6)
Message boards :
RALPH@home bug list :
Report - Previously Unclassified Work Unit Errors
(Message 821)
Posted 6 Mar 2006 by Spare_Cycles Post: Carlos, I think the most probable explanation for SIGSEGV is because your Linux PC has only 256MB of RAM, whereas your WinXP PC has 512MB RAM. The lack of physical memory will never cause a SIGSEGV on a properly functioning modern PC. Programs run in virtual memory, and the virtual memory will look exactly the same regardless of how much physical memory there is. If there isn't enough physical memory then there will be a lot of swapping to disk, which can slow things way down. That can cause a problem if the computer is doing something like burning a CD. It will never cause an error in a crunching program like ralph/rosetta. That assumes the PC is working. If, for instance, there are errors when reading the hard disk, then pages will be corrupted when they are swapped back in. |
|
7)
Message boards :
RALPH@home bug list :
Report - Previously Unclassified Work Unit Errors
(Message 811)
Posted 3 Mar 2006 by Spare_Cycles Post: This WU finished using v 4.92 and claimed credit but contained some interesting messages so worth a look by the experts: Looks like the WU errored out and you would have gotten zero credit, but the new code that we're now testing kicked in and salvaged the WU. |
|
8)
Message boards :
RALPH@home bug list :
Report - Previously Unclassified Work Unit Errors
(Message 759)
Posted 1 Mar 2006 by Spare_Cycles Post: build with original Libc.6.so -or- newer libc.6.so -:) My gentoo installs would be the newer libc.6.so. The "2.4.20-4GB-athlon" install is SuSE 8.2 and most of the libs are dated march 13 2003 and seem to be version 2.3.2. I don't know why these would work but not yours. |
|
9)
Message boards :
RALPH@home bug list :
Report - Previously Unclassified Work Unit Errors
(Message 747)
Posted 28 Feb 2006 by Spare_Cycles Post: this message appeared on my remote ssh Linux console´ Have you checked this computer with programs like memtest86 and Super Pi? may be the purpose is disallowing who owns a Linux Pc w/ Kernel 2.4.x from I have five Linux PC Kernel 2.4.x computers running RALPH. So far they have completed all WUs successfully except for a few that nobody seems able to crunch. |
|
10)
Message boards :
Feedback :
Difference F@H and R@H???
(Message 687)
Posted 26 Feb 2006 by Spare_Cycles Post: F@H uses a low-level but physically accruate method to simulate the folding of a protein. This requires a huge amount of CPU time, like 10,000 WUs for a fast-folding protein. Because it's physically accurate, it gives a good idea of how the protein actually goes about folding. It is easy to get the sequence of a protein, but hard to find it's structure. With the F@H method it is possible to deduce the structure from the sequence, but only for fast-folding proteins. For most proteins of interest, the F@H method is too slow. Rosetta tries to use a higher-level approach to fold the protein. This is potentially much faster. Instead of taking 10,000 WUs for a single fold, rosetta can do maybe 100 folds in a single WU. The trouble is, the answer rosetta gets isn't all that great. Rosetta could get a much better answer with some more work on the algorithm. That will take a fair amount of CPU power, though, as many ideas will need to be tested under a number of conditions. And that is what rosetta is about. We provide the CPU power, and the development team uses this power to improve the algorithm. |
|
11)
Message boards :
Feedback :
Difference F@H and R@H???
(Message 685)
Posted 26 Feb 2006 by Spare_Cycles Post: a dumb waste of computing time... I smell a troll. |
|
12)
Message boards :
Cafe RALPH :
(DO NOT POST HERE) This is the Moderators Archive thread
(Message 646)
Posted 25 Feb 2006 by Spare_Cycles Post: Some want red, some want black, some want bold, so maybe everyone will be equally dissatisfied with red-black-bold. :) (Or maybe you could threaten to use cyan. :p ) |
|
13)
Message boards :
Feedback :
Credit scores
(Message 412)
Posted 21 Feb 2006 by Spare_Cycles Post: The credits system is done by BOINC not Baker lab. So why would you not bring the issue to BOINC instead? BOINC provides a credit system, but it works really badly for a no-redundancy project. And as BOINC is open source, this probably can't be fixed. A project doesn't have to use the BOINC system, though, it can use some method of it's own to estimate the amount of work done. For instance, the project's client could count the number of times certain low-level subroutines are called and base its estimate of work done on that. Such a solution would presumably be implemented and tested with RALPH, then moved to ROSETTA once it was ready. |
|
14)
Message boards :
Feedback :
Credit scores
(Message 403)
Posted 21 Feb 2006 by Spare_Cycles Post:
I don't understand why this isn't a good place for FluffyChicken to post this. The change he is requesting would need to be implemented and tested in RALPH. Personally, I agree with FluffyChicken. The current points system is so badly flawed that it takes the fun out of competing. People who enjoy the points competition have to go somewhere besides rosetta to find a level playing field. |
|
15)
Message boards :
Number crunching :
No work from project.
(Message 234)
Posted 18 Feb 2006 by Spare_Cycles Post: I'm running it split 50/50 with rosetta@home, will it automatically keep returning to rosetta units till it finds some work from Ralph? I'd hate my PC to sit doing nothing just because there is no work available from one of the projects. BOINC will run whatever projects you have attached for which there is work. BOINC will keep track of the fact that it is running one project much more than you have specified, and once the down project comes back BOINC will run that for a while until it catches up. Thus, over the long term BOINC will split up the computers time the way you asked. |
|
16)
Message boards :
Number crunching :
Need for Linux testing?
(Message 232)
Posted 18 Feb 2006 by Spare_Cycles Post: good, however all are kernel 2.6.x none of yours has kernel 2.4.x The computer I have on ralph uses SuSE Linux kernel 2.4.20-4GB-athlon. I haven't had any problems except for not being able to get work (only 2 WUs so far, both successfully crunched). On rosetta I have 10 computers using Linux kernel 2.4.31-gentoo-r1. I would estimate that they fail to successfully crunch a WU less than one time in 1000 (ignoring the WUs that nobody can crunch, and a brief period when all machines got download errors when trying to get work.) |
|
17)
Message boards :
RALPH@home bug list :
Registering with boinc account to RALPH
(Message 99)
Posted 17 Feb 2006 by Spare_Cycles Post: I tried signing up with a new email and it didn't work. Eventually I manually created the account, and then I was able sign up with the (now existing) account. |
©2024 University of Washington
http://www.bakerlab.org