Caputring WU stats
11 Dec 2005 1:31:07 UTC
Topic 190348
(moderation:
I am hoping someone can tell me if it is possible to progarmaticly capture the following information about a WU. When was it initially download, when was it terminated or uploaded, cpu time, claimed credit, when was the status determined, end status, granted credit. Based on what I have seen I think the answer is yes but have found nothing that explains how to go about it. I love running statistic on such data and have been gathering some crude daily stats but would like better data to work with.
I have found some of the information in the files used to generate the graphs for the boinc statistics tab but other components are missing.
Caputring WU stats
)
If you are on Windows, BOINCView does a lot of this for you. Otherwise, all I can tell you is that it's all in the various xml files, you would have to capture them every "x" minutes...
Bill Ok, I can handle that.
)
Bill
Ok, I can handle that. Does the info include server side updates on each WU?
RE: Does the info include
)
I've only played with BOINCView a bit, don't normally use it, but I don't think it has any server-side info. For that, you'd have to look at something like Andy K has done at http://www.andyk.de/boinc/pc.php?pid=eah&uid=72129&hostNum=1&hostID0=&ev=on&av_seconds=on&avcc=on&avgc=on (id plugged in there is yours, just for an example).
Edit:: Corrected URL, but still having connection problems to his site right now...
Bill Thank you, this looks
)
Bill
Thank you, this looks very close to what I want. And Andy is even offer the soruce code!
RE: this looks very close
)
You must have made it through when I couldn't, and figured out what I messed up in the URL, because you posted before I edited... Good work! :-)
Be aware that it can only track what is still on each project's site - and the projects DO delete old data every couple of weeks or so... so you'll need to keep a local copy of any data.
Bill I understand about
)
Bill
I understand about deleting the data, but since they keep the data for a week after the last update so daily queries should be sufficent. I supect there will be a lot of parsing and sorting but I am good with that kind of code. Making something that is easy for others to understand and use I am kind of weak at.