Page 1 of 1

flops/day?

Posted: Sun Dec 15, 2013 3:45 pm
by trench
Hi!

I just started folding again (after a break for several years) and are now building a GTX Titan rig to support this cause. I'm wondering if there's a chart somewhere that shows the number of folding CPUs over time? Also the total number of points per day over time would also be interesting - I imagine that this has increased plentyfold since the v7 client. Aggregated for all workers, per team or client, any number would be nice to check out :)

Thank you for keeping up the great work you're doing for all of us and helping curing major human diseases!

Re: flops/day?

Posted: Sun Dec 15, 2013 4:20 pm
by Mike_Shaffer
trench,

Welcome, from one new forum member to another. The current flops info is at http://fah-web.stanford.edu/cgi-bin/mai ... e=osstats2 but not sure of where that would be charted/tabulated.

Re: flops/day?

Posted: Sun Dec 15, 2013 4:22 pm
by Mike_Shaffer

Re: flops/day?

Posted: Sun Dec 15, 2013 5:20 pm
by trench
Thank you for the links! I'm now checking out extremeoverclocking.com and hope that their database contains the historical numbers I'm interested in. It would be cool to see the computing power development over time - it must have increased multiple times just because of the new code in v7- that takes advantage of the new GPUs.

Re: flops/day?

Posted: Sun Dec 15, 2013 5:27 pm
by Mike_Shaffer
trench,

I think it is related to the increase in cores and processing capability of the 16, 32, 48, CPU core folders running v6 clients with 2P or 4P configurations and with certain options set, like bidadv. Also, V7 and the new video cards, like the Titan which is one of my full time folders, have 2000+ cores.

Re: flops/day?

Posted: Sun Dec 15, 2013 5:46 pm
by trench
Cool, I guess that would be a nice dimension to the chart - were all the computing power comes from. If the data exists (and hopefully it does thanks to extremeoverclocking.com) I guess a stacked line chart would visualize this in a nice way. I'll contact the database owner and continue my dig :) I'll update this thread with my progress.

Re: flops/day?

Posted: Sun Dec 15, 2013 6:57 pm
by Grandpa_01
I am amusing this would be in line of what you are looking for although the quit keeping data last year as near as I can tell. https://docs.google.com/spreadsheet/ccc ... pY1E#gid=0

Re: flops/day?

Posted: Sun Dec 15, 2013 7:33 pm
by Nathan_P
Gramps, here's the current one:-

https://docs.google.com/spreadsheet/ccc ... CdVE#gid=0

Re: flops/day?

Posted: Sun Dec 15, 2013 7:34 pm
by Jesse_V
Grandpa_01 wrote:I am amusing this would be in line of what you are looking for although the quit keeping data last year as near as I can tell. https://docs.google.com/spreadsheet/ccc ... pY1E#gid=0
That's because we moved to a new chart: https://docs.google.com/spreadsheet/ccc ... ring#gid=0

To make things easier, I made a chart of all available data and plotted it with three fastest supercomputers from each Top500 list:
This was the result: https://commons.wikimedia.org/wiki/File ... Powers.png
Image

Re: flops/day?

Posted: Mon Dec 16, 2013 12:56 pm
by trench
Wow! Just looking at what has happened since 2012 - capacity has _doubled_ - IN A YEAR (!) This project is truly a fantastic way to contribute to research and an innovative way to utilize all unused computing power. Thanks!

Re: flops/day?

Posted: Tue Jan 28, 2014 11:18 pm
by 7im
Jesse_V wrote:...

That's because we moved to a new chart: https://docs.google.com/spreadsheet/ccc ... ring#gid=0
Just got a chance to look at the new spreadsheet in detail.

Would it be possible to pull some of the data from the older spreadsheet to the newer spread sheet? You wouldn't have to go back all 10 years, but the last 1 or 2 years pulled over for the charts would give the current numbers better perspective, IMO. It would also show a better (or lessor) continuity in the participation numbers.

And the new charts are divided by year, not continuous. So again, the charts lose scale and reference to other years.

Statistics

Posted: Tue Mar 04, 2014 12:40 pm
by trench
Hi!

I started collecting statistics from FAH in the end of 2013. I've created a simple page that shows global Folding@Home statistics (over time). Please check it out, and feel free to send questions & feedback: http://app.johanssonrobotics.com/folding/

Have a great day all!

Re: Statistics

Posted: Tue Mar 04, 2014 2:10 pm
by Kurtis200200
Actually pretty excellent; surprising that so much of the computing power appears to come from older Fermi GPUs

Re: flops/day?

Posted: Tue Mar 04, 2014 10:32 pm
by Joe_H
The Fermi GPU group includes Kepler and later nVidia GPU's as well. The nVidia classification is all of the models prior to Fermi as I understand the groups.