Folding@Home Benchmarking Thread

WYP

News Guru
Did you know that there is a dedicated Folding@home benchmarking tool to see how well your GPU will fold?

We all love benchmarking here so I thought I would raise some awareness for this benchmarking tool and the Redline@OC3D Folding Team by starting a little Benchmark thread.

To take part all you will need to do is download version 2.2.5 of the FAHBench and test the single and double prescision performance of your GPU. For more accuracy please use a run of 120 seconds.

https://fahbench.github.io/

Please post screenshots of your results like below. I can then place your result in the scoreboard below.

We also need to know your GPU clock speeds if you run an overclock.

wTcm21f.jpg

vv9u0d5.jpg



Single Precision Performance

1. QuietOne, EVGA GTX 980Ti classified @1529MHz , 105.386
2. QuietOne, EVGA GTX 980Ti classified, 103.396
2. Kilbane, GTX 970 Watercooled @ 1503MHz, 78.0339
3. Wyp, R9 Fury X @ Stock, 73.391
4. Kilbane, GTX 970 Watercooled @ Stock, 73.0365
5. Jimma47, ASUS GTX 980 Strix @stock, 72.2604
6. Tolemac, Sapphire R9 290X @1030MHz, 64.1218
7. Thelosouvlakia, R9 290 @ Stock, 52.8489
8. Wyp, R9 380 @ stock, 45.0353
9. hmmblah, GTX 960, 44.7992
10. Wraith, KFA GTX 950 EXOC @ Stock, 44.6526
11. Wyp, Intel i7 6700K @4.7GHz, 3.1196
12. Wraith, Intel i7 4790K 4.0GHz CPU, 2.40781

Double Precision Performance

1. Tolemac, Sapphire R9 290X @1030MHz, 14.7746
2. Thelosouvlakia, R9 290 @ Stock, 13.0258
3. Wyp, R9 Fury X @ Stock, 11.4879
4. QuietOne, EVGA GTX 980Ti classified @1529MHz , 8.46982
5. Jimma47, ASUS GTX 980 Strix @stock, 5.37954
6. Wyp, R9 380 @ stock, 5.51127
7. Kilbane, GTX 970 Watercooled @ 1503MHz, 5.19795
8. Kilbane, GTX 970 Watercooled @ stock, 4.71508
9. hmmblah, GTX 960, 3.25077
10. Wyp, Intel i7 6700K @4.7GHz, 3.11424
11. Wraith, KFA GTX 950 EXOC @ Stock, 2.69415
12. QuietOne, Intel i7 4790K 4.0GHz CPU, 2.50324
13. Wraith, Intel i7 4790K 4.0GHz CPU, 2.41974
 
Last edited:
Pretty neat, hadn't heard of this until now. Here's my 960 results, I fold on two 960's. I'll try to do the 980 tonight.

v3Uss0y.png

bkrFPKu.png


Edit:

GPU-z (stock)

bd3TMoZ.png
 
Last edited:
All added to the OP. Very interesting results.


Question to any folders out there. How important is double precision performance for folding?
 
New results have been added to the OP. Thanks for the results so far. Some interesting results.
 
Intresting things happening.

Both are 980 Strix running as stock as they came out of the box.
Single Point - Never got above 72% usage. Double was closer to 98%

Running dual cards one was slower in both tests but I don't think it would be due to throttling etc as neither really broke a sweat as far as memory usage, heat or voltage.
I'm not sure how much of a difference it is in real life.

Single Point

mzm08mq.png


Double Point

jckSRGF.jpg
 
OP updated with new results.

AMD is just raking it in with the double precision performance. Interesting how hawaii is better than Fiji.
 
OP updated with new results.

AMD is just raking it in with the double precision performance. Interesting how hawaii is better than Fiji.
AMD has always had the edge in OpenCL with Double floating point precision, it's single point they struggled with.. roll on Vega and Polaris.
 
List updated.

Do any of you know how often Double precision loads come up for F@H? I've read that 90% of work units are Single, but that data is really old.
 
Last edited:
Back
Top