Hi FAH team,
I've been reading a paper written by the PS3GRID group, another scientific group that tries to do science on the
PS3 and they speak of 'the extremely volatile on-grid-persistence of the average user', can de FAH team, as you must
no doubt have vast statistics on the subject, say anything about this based on your own experience with this project,
are we (contributors) also really so unreliable? An then more interestingly, can the award system be adapted to give
reliable clients (always return their WU's in time) more points?
greetings,
Harm ten Napel
Reliability and Trust of the FAH contributors
Moderators: Site Moderators, FAHC Science Team
-
jrweiss
- Posts: 704
- Joined: Tue Dec 04, 2007 6:56 am
- Hardware configuration: Ryzen 7 5700G, 22.40.46 VGA driver; 32GB G-Skill Trident DDR4-3200; Samsung 860EVO 1TB Boot SSD; VelociRaptor 1TB; MSI GTX 1050ti, 551.23 studio driver; BeQuiet FM 550 PSU; Lian Li PC-9F; Win11Pro-64, F@H 8.3.5.
[Suspended] Ryzen 7 3700X, MSI X570MPG, 32GB G-Skill Trident Z DDR4-3600; Corsair MP600 M.2 PCIe Gen4 Boot, Samsung 840EVO-250 SSDs; VelociRaptor 1TB, Raptor 150; MSI GTX 1050ti, 526.98 driver; Kingwin Stryker 500 PSU; Lian Li PC-K7B. Win10Pro-64, F@H 8.3.5. - Location: @Home
- Contact:
Re: Reliability and Trust of the FAH contributors
Do you have a public link to the paper available? It would be interesting to see the author's credentials and the context of his statements...
I suspect that volatility in the contributor base is a 'fact of life' in any large, voluntary Distributed Computing project. Whether or not that volatility translates to "unreliable" may be a matter of opinion and project-specific factors.
As for F@H, "reliable" contributors already get more points than those who do not return WUs in time. If you turn in a WU after the final deadline, you get 0 points...
There have also been proposals and discussions to adjust the award system such that those who turn in WUs even earlier will get a points bonus. One proposal was to tie such a bonus to the PF (performance factor) already calculated by the Project. When and how any such adjustment of the awards system will happen is up to the Pande Group.
I suspect that volatility in the contributor base is a 'fact of life' in any large, voluntary Distributed Computing project. Whether or not that volatility translates to "unreliable" may be a matter of opinion and project-specific factors.
As for F@H, "reliable" contributors already get more points than those who do not return WUs in time. If you turn in a WU after the final deadline, you get 0 points...
There have also been proposals and discussions to adjust the award system such that those who turn in WUs even earlier will get a points bonus. One proposal was to tie such a bonus to the PF (performance factor) already calculated by the Project. When and how any such adjustment of the awards system will happen is up to the Pande Group.
Ryzen 7 5700G, 22.40.46 VGA driver; MSI GTX 1050ti, 551.23 studio driver
Ryzen 7 3700X; MSI GTX 1050ti, 551.23 studio driver [Suspended]
Ryzen 7 3700X; MSI GTX 1050ti, 551.23 studio driver [Suspended]
-
VijayPande
- Pande Group Member
- Posts: 2058
- Joined: Fri Nov 30, 2007 6:25 am
- Location: Stanford
Re: Reliability and Trust of the FAH contributors
I think they're talking about the fact that donors can come and go, even before a given WU is completed. To handle all of this smoothly, one needs well thought out server side algorithms. FAH's been working on them for almost a decade and they've gotten pretty sophisticated, especially in terms of how we can combine lots of trajectories (which may or may not be returned) into useful results. Much of those ideas are detailed in our papers, but it takes a lot of work to implement all of those ideas.