[TheBlademaster01]

As you might have noticed work units fluctuate quite a bit in PPD output, so depending on the unit you are folding you might get highly different results. This database was created to compare frame times with other users in order to see whether your PPD is where it should be or if there is something wrong with the system (instability, incorrect client configuration, throttling etc.)

Fill out the form below to submit data to the database and please maintain the proper format as good as possible.

(Form For Submission | Full Page Spreadsheet)

https://docs.google.com/spreadsheet/pub?key=0AmQ6YZGMTUpFdHhxNkptVncyR2k4dVV6WDBmQXRJYVE&output=html&widget=true

Legacy Database

 

[TheBlademaster01]

Quote:

Originally Posted by Erick Silver

Info added.

First Radeon results

Finally a new GPU eh?

Quote:

Originally Posted by msgclb

same same

Thanks

 

[TheBlademaster01]

Oh, I'm going to let you do your thing anubis lol

 

[TheBlademaster01]

Quote:

Originally Posted by Erick Silver

Yep, a freebie at that. Its been my bestest freind eber!!!

I fold 24/7 on the HD7950 and game on the GTX560SE for now. I am hoping to be able to upgrade my PSU and then purchase a HD7970/7990 to fold on 24/7 and game on the 7950 at some point. Right now though, I have to move, so no new parts.

Wow, that's cool. I got a couple of freebies as well over the summer (a package to be accurate though

). Still haven't had the opportunity to put it all to good use

 

[TheBlademaster01]

Yeah, anubis said the PRCG mattered for PPD so I included it. It's should be listed in the History viewer though (rightmost columns).

 

[TheBlademaster01]

Well, I think it's informative though. It seems the newer runs are more efficient from these results.

 

[TheBlademaster01]

I think a 660Ti should be cheaper and produces more. As for a GTX 760, I'd imagine somewhere between 50-75k. Sorry, I don't have too much data on this

E:

ninja'd by 'nub...

 

[TheBlademaster01]

pain and pills???

What happned?

 

[TheBlademaster01]

Ah, that sucks man

. Not enough exercise before? I get weird issues with muscles etc when I get rusty from not moving enough

Of course things can just happen to you as well.

 

[TheBlademaster01]

Yeah I think HFM has an issue with your time zone I assume GMT+3 (12:51 where you are). It just increases wall clock by 3hrs but is not accurate for PPD calculations.

Frame times are not based on wall clock and will give you a good approximation if you fold continuously and have consistent frame times. It will however have a slight skew due to load imbalances, upload times and possible pauses.

Also, just a note. You don't have to upload all your results to the database since it could get tedious to read rather soon. Unless you get a weird result, OC'ed the GPU some more (or underclocked

) or another variable. Thanks for the comtributions though

 

[TheBlademaster01]

Tanks gents

 

[TheBlademaster01]

You can submit it using the form

Also, you need to provide data on TPF, OS and mem clocks.

 

[TheBlademaster01]

It's memory clock, and no the quotation marks aren't needed. Maybe need to add that notion later...

Memory clock should be listed in GPU-Z or MSI afterburner etc

If you have GDDR5, multiply the number by 4 for the effective frequency.

 

[TheBlademaster01]

It's effective meaning for GDDR4 that it transfers 4x per clock cycle.

Keplers run at 1502 mostly, which means 6008 effectively.

 

[TheBlademaster01]

Thanks Sam

 

[TheBlademaster01]

GTX 780Ti = 5.05 TFlops
GTX 780 = 3.98 TFlops

A GTX 780Ti should be about 25-30% faster based on TPF. QRB should inflate this difference some more.

 

[TheBlademaster01]

About the same difference as a 680 and a 780 I think. 780 indeed has the better price/perf

 

[TheBlademaster01]

Lots of particle dynamics/lighting effects are run off the GPU as well now. Also, higher resolutions, higher quality/larger textures and supersampling require more fragments to be rendered and converted into pixels. This raises the need for more ROPs, TMUs and bandwidth to keep up with the data traffic.

Often with CG the computational expense of some of these effects is far larger than you would visually estimate/appreciate. 2x2 SSAA takes exponentially more time to render vs something like no AA or coverage (CSAA/EQAA) or even post process edge (FXAA/MLAA) based blurring. But it might not give an exponentially better result. An already good result is difficult to improve on but can be done with huge enough advancements and displays --think high quality/frequency monitors or in case of audio better speakers. Then there is the whole gaming console story

.

There are also things like GPGPU (simulations, ray tracing etc.) that see (near) linear performance increases. Try ray tracing with a 9800GTX vs 780Ti or just look at the difference in folding performance despite the 9800GTX running on an optimized API like CUDA vs 780 on OpenCL (granted the 780 gets Quick Return Bonus).

The advancements are necessary but it depends on your personal use cases whether or not you'll need /notice them

 

[TheBlademaster01]

E2200 is a E4500 with it's L2 stripped in half. Which was a midrange Core 2 Duo in 2007. I still have mine, had it overclocked to 3.7GHz once, though 3.1GHz for 24/7

Yes it can pull 2 GPU clients.

 

[TheBlademaster01]

New batch processed.

Arvid is ninja'ing some crazy results. P9401 seem to be the new hype lol. Also congrats on being the new PPD king at 320k PPD