Posts by kksplace

1) Message boards : Number crunching : New jobs in Rosetta Python projects (Message 108891) Posted 1 Mar 2024 by kksplace Post: Well, I was just letting BOINC do its scheduling thing. With my particular settings, up to 4 at a time were running. I looked up how to limit them (max_concurrent in app_config) and am now trying limiting to two at a time. Just got the first two about an hour ago, so we will see in about 7 hours....
2) Message boards : Number crunching : New jobs in Rosetta Python projects (Message 108889) Posted 1 Mar 2024 by kksplace Post: I added the VM Extension pack, but still having problems. The WUs will now at least run for a while (hours in most cases) but then still error out. Anyone figured out what else to do?
3) Message boards : Number crunching : New jobs in Rosetta Python projects (Message 108794) Posted 19 Jan 2024 by kksplace Post: You asked not to abort these tasks. I received four today. They all show "Postponed: Communication with VM Hypervisor failed." after about 12 seconds. Should I just leave them or abort them?
4) Message boards : Number crunching : Linux tasks (Message 96819) Posted 28 May 2020 by kksplace Post: I agree with you that credits don't mean much of anything on their own, especially comparing projects. The only point here (that I was not clear about) is that within Rosetta, it is about the only way to compare between CPUs, operating systems, etc. It is interesting that you see less credits between Linux and Windows. One of my setups is a a dual boot Linux or Windows, and it definitely shows better credits when on Linux vs. Windows, and not only on Rosetta. Happy crunching!
5) Message boards : Number crunching : Linux tasks (Message 96799) Posted 26 May 2020 by kksplace Post: Tullio, be aware that for Rosetta the time a work unit runs is chosen by you. The default is 8 hours. As mentioned below, you can change the default time shorter (or longer). Credits are based on how much 'stuff' your computer/CPU calculates within that time frame. So you will not see very much variation in run times. (Occasionally one will run a shorter time when some 'maximum' amount of decoys is reached, or it may run up to a certain time over your chosen time attempting to accomplish a minimum amount of decoys.) So the real comparison should be between average credits given for your chosen time-frame...and it should be an average; credits vary from WU to WU depending what was accomplished. I have attempted to follow the separate threads on how credits are given, but quite frankly don't understand the details behind the science. In general, the more science done during a given time for a given WU, the more credit.
6) Message boards : Number crunching : Discussion of the merits and challenges of using GPUs (Message 96727) Posted 22 May 2020 by kksplace Post: GPUGrid again has workunits available.
7) Message boards : Number crunching : Rosetta 4.1+ and 4.2+ (Message 94445) Posted 14 Apr 2020 by kksplace Post: Will creating the cc_config.xml file input that is recommended also limit Rosetta Mini WUs that are also i686? They seem to be working fine (actually working quite well based on RAC) on my Linux rigs. It is only the 4.12 and now 4.15 i686 WUs that are having problems. (...and I got signed up on Ralph, awaiting test WUs when they are available for one of my Linux rigs.)