Posts by Ray
|
1)
(Message 5735)
Posted 11 Mar 2018 by Ray Post: Just handwaving here, but... There are (ATM) 222839 WUs out, and 36900 computers with recent credit. That works out to almost exactly 6 WUs per recently active computer... for my GPU, that's a bit over 5 hours of work, and my GPU isn't that special. Looking at it another way, since a WU is 480 credits, there are just under 107 million credits of work outstanding... since BoincStats shows 65 million credits of work reported yesterday (for this project), that's less than a day and a half of work at that rate (and on the 9th, 129 million credits of work were done, which the highest in the last 40 days, and ~3x what the low was). Seems obvious that 3 days is too long for this project the way it's currently set up... when a new batch of work is added, the people that ping the server first suck it all up, even without 'irrational' settings.... IIRC, the default setting for the BOINC client is to grab 10 days of work. |
|
2)
(Message 5732)
Posted 10 Mar 2018 by Ray Post: It's most likely a Windows setting (not a BOINC setting) that is putting your computer to sleep. Windows only thinks the computer is being 'used' when there is user input. Look at your "Power & Sleep" settings in Control Panel. |
|
3)
(Message 5731)
Posted 10 Mar 2018 by Ray Post: WUs are becoming available, but slowly... if you look at the server status, there are nearly a quarter of a million WUs out for processing, and all of the ones I'm getting are for 'validation'. If you don't connect to the server very often, other people grab them up before you. It looks like there are a lot of people out there who are 'holding on to' very large queues of first-run WUs, and starving out everyone else. |
|
4)
(Message 5730)
Posted 10 Mar 2018 by Ray Post: For what it's worth (I've only recently rejoined this particular project) but I am slowly getting a trickle of 'validation' WUs, though my computer pings the server fairly often. It looks like people are just grabbing the returned ones for validation much faster than people are processing the ones currently out.... BoincStats shows 64 million work credits earned on this project yesterday, so work is clearly getting done. |
|
5)
(Message 5729)
Posted 10 Mar 2018 by Ray Post: Um, no. See https://www.sciencedirect.com/science/article/pii/S0019103517300313. Dr. Durech, the lead author of that article, is the project director. The peer review process for publication is really slow... that article was submitted to arXiv in July of last year, and it's only hitting the press next month. They are clearly still working in the field, and while that paper doesn't specifically cite this project, from reading the preprint it's pretty clearly the same work. |