Posts by [HWU] GHz

1) Message boards : Number crunching : Problems with Rosetta version 5.67 (Message 41549) Posted 28 May 2007 by [HWU] GHz Post: I also have memory problems with gp04* WU: http://boinc.bakerlab.org/rosetta/result.php?resultid=82333923 http://boinc.bakerlab.org/rosetta/result.php?resultid=82389538 http://boinc.bakerlab.org/rosetta/result.php?resultid=82411587 http://boinc.bakerlab.org/rosetta/result.php?resultid=82433422 This wu crashed at the start of computation.
2) Message boards : Number crunching : Set quorum to 2 WU to avoid cheating (Message 7758) Posted 27 Dec 2005 by [HWU] GHz Post: This has already been discussed Sorry, i did not see any discussion about this - the benchmarks on that machine are wrong. There is no evidence of cheating. Do you thik that is a boinc error? This user's data will have to be adjusted after the project staff returns from the holidays. Ok :) Setting the quorum to 2 (or redundancy in general) has been discussed as well. Unless the _current_ plans (I believe) to replace the benchmarks with flops-counting fall through, there is little or no need to waste the computing resources on redundancy. Flops-counting? Where I can find information about it? The occasional problem like this one host can be dealt with manually. until the problem is only one user there are no problems...;) Bye, GHz
3) Message boards : Number crunching : Set quorum to 2 WU to avoid cheating (Message 7750) Posted 27 Dec 2005 by [HWU] GHz Post: Do you have see the new first user of rosetta? http://boinc.bakerlab.org/rosetta/top_users.php And its host: http://boinc.bakerlab.org/rosetta/top_hosts.php http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=105526 And this is the onli assigned WU: http://boinc.bakerlab.org/rosetta/workunit.php?wuid=3530614 Looking at the benchmark, i think that joseph is cheating and with the quorum of 1 WU it's very easy to do. Please set the quorum to 2 WU and reset the total credit for that user and its host. Bye, GHz