I'm Folding!

[GodSpoon]

I started yesterday, my main computer is folding whenever I'm not using it and I decided to run a Chrome WU on my laptop for the fun of it

 

My main rig is a cute G3258 @ 4.8Ghz and an R9 280x Black Edition

 

Is there a good way to measure PPD metrics? Or is it irrelevant due to different WUs?

 

I also tried folding on my 2014 mbp on Chrome, cranked out a WU in 12 minutes 

 

I know those are generally smaller WUs in the chrome client but still impressive for a laptop!

 

But yeah! Currently on my 12th WU! 

 

Thank you to people like @Whaler_99 and others who explained in great detail what folding is and how to do it! 

 

Im glad to be helping the cause

 

Edit: I see that PPD metric in the client now I was looking at the chrome client on my laptop at the time, my desktop is supposedly getting 158k PPD

[TheRandomness]

Every bit helps c: Also, it should say on the client on what PPD you should be getting, although let this number refresh for the next few hours to get an accurate number.

[royalcrown]

IMO it's not THAT exciting...lol. Still good to do though. 

[SansVarnic]

        :lol:

[tobben]

Welcome to the team dude, fold on!

 

IMO it's not THAT exciting...lol. Still good to do though. 

But that is exactly what it is, it's THAT exciting!

[GnatDaGnat]

Ya, it is a bit exciting at first.  Good to do, and I just started folding on Linus's team today.  Just earned the team 207 points.    Glad to see more than just me starting!

 

Wait Linus has a team?? Number pls?

[petem]

Welcome to team Linus!

 

That's a huge overclock  .  If you want to determine the relative folding power of your GPU, snag a copy of FAHbench from https://folding.stanford.edu/home/download-utilities/

It'll generate 4 metrics for single and double precision, explicit and implicit computational scores (in ns/day-nanoseconds per day), which you can use to compare to other gpus.

 

Per the Stanford page "The single precision implicit model most closely relates to current folding performance.", but that page was last updated Oct. 2014.  If single precision implicit computational modeling is still the predominant task, then that would be the best comparison between GPU card performance.  

 

Work Units belong to different "cores"; some of them use explicit solvation calculations to include the surrounding solvent (e.g. water), while others use implicit solvation which treats the solvent as a mathematical continuum.  I haven't done an exhaustive search, but so far I haven't found a definitive guide to what is the most common type.  Absent information contradicting the Stanford webpage, I use the single precision implicit score for comparison purposes.  

 

Their is an outdated table of FAHBench scores at https://foldingforum.org/viewtopic.php?p=242081, and Anandtech seems to be the best GPU reviewing website that includes "compute" FAHBench scores, but they don't have a comprehensive set of GPU reviews.  I noticed their GTX 970 review from 9/26/2014 includes an entry for a generic R9 280x in their FAHBench charts.  The single precision implicit score was 146.6, which is pretty good for a 2 year old card, and definitely better than my 3 year old GTX 660ti!

 

I'm surprised there isn't a benchmarking website that gathers and tabulates that info, but I have only seen partial or outdated attempts to do so, and usually buried in a forum somewhere.  

Anyhoo, that is a discussion for another day (or at least another thread   )

[GangstaRas]

I'm surprised there isn't a benchmarking website that gathers and tabulates that info, but I have only seen partial or outdated attempts to do so, and usually buried in a forum somewhere.  

Anyhoo, that is a discussion for another day (or at least another thread   )

 

Like this one ? https://docs.google.com/spreadsheets/d/1vcVoSVtamcoGj5sFfvKF_XlvuviWWveJIg_iZ8U2bf0/pub?output=html

Like you OP I pretty much started about a week ago, I have about 17 WU going 18 in the next 22 minutes as I type here. I think your PPD is going to increase overtime, as when I started, I had like 43k PPD to my GTX 960 and thought that was good because I was also using my i3 3240 to do some work of its own at 6k PPD. It was when I did some research that I realized I shouldn't have these kinds of speeds, but what it seems like to me is that its something that matures the more you fold. I watched the progression of my PPD move from 43k to 80k then to 166k and now I'm around 173k without doing a thing but do more of them, and based on this database at least, it would seem that my peak is going to be 188k, and theoretically 200k for yours, so I'll say keep an eye on it

[petem]

Like this one ? https://docs.google.com/spreadsheets/d/1vcVoSVtamcoGj5sFfvKF_XlvuviWWveJIg_iZ8U2bf0/pub?output=html

Like you OP I pretty much started about a week ago, I have about 17 WU going 18 in the next 22 minutes as I type here. I think your PPD is going to increase overtime, as when I started, I had like 43k PPD to my GTX 960 and thought that was good because I was also using my i3 3240 to do some work of its own at 6k PPD. It was when I did some research that I realized I shouldn't have these kinds of speeds, but what it seems like to me is that its something that matures the more you fold. I watched the progression of my PPD move from 43k to 80k then to 166k and now I'm around 173k without doing a thing but do more of them, and based on this database at least, it would seem that my peak is going to be 188k, and theoretically 200k for yours, so I'll say keep an eye on it

 

Oooo!  I saw that - but I was hoping for a power usage column too - hard to compare apples to apples when you're trying to determine the best way to go, especially considering power usage during folding doesn't match "gaming" "average" or "peak" usages commonly posted in reviews... but none the less, that's a good start