Page 6 of 17

Re: PPD Database - HFM exports needed

Posted: Wed Apr 29, 2020 5:31 pm
by foldinghomealone
HaloJones wrote:You could merge the results for the "RX 5700XT" and the "RX 5700 XT"
Thanks, done.

I also deleted a few entries where PPD was way too low compared to average

Re: PPD Database - HFM exports needed

Posted: Wed Apr 29, 2020 5:49 pm
by NormalDiffusion
Endgame124 wrote:If I can figure out remote management, I can provide folding scores for:

APU (both CPU and integrated video)
A10-5800K
A10-7870K

Cpu:
Ryzen 7 2700X

GPU
R250X
Geforce 1080ti.

Is any compelling reason for me to collect any of those? If so, which ones?
Do it old-school like me! Walk to the machines! :)
foldinghomealone wrote:I think 1080ti and R7 2700X are commonly used but under represented in the DB.
I would be happy to see some more data
+1 for the 1080ti!

Re: PPD Database - HFM exports needed

Posted: Wed Apr 29, 2020 5:51 pm
by NormalDiffusion
foldinghomealone wrote: I also deleted a few entries where PPD was way too low compared to average
I think there is a bug in hfm.net. Saw it also on a couple of GPUs where hfm was all over the place...

Re: PPD Database - HFM exports needed

Posted: Wed Apr 29, 2020 6:03 pm
by Endgame124
We could really use some 2060 and 2060 super numbers. It would really help with the points per watt purchasing decisions, especially when the next gen cards land

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 1:24 am
by NoMoreQuarantine
For fun, I used the data to create some graphs showing performance across frame sizes (GPUs only for now)

Unfortunately, we're still lacking samples for a lot of processors. I wish we had a tool that could quickly feed the processors frames of various sizes and measure their performance.

Edit: Just realized that my analysis was faulty. Deleted the link to the graphs until I can fix them.

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 3:08 pm
by foldinghomealone
New updates for
GTX 1660 Ti
RX 560
R5 3600

Thanks to
NoMoreQuarantine
DmitryKo

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 4:26 pm
by NormalDiffusion
NoMoreQuarantine wrote:
Edit: Just realized that my analysis was faulty. Deleted the link to the graphs until I can fix them.
What was wrong? Was having a quick look this afternoon...

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 5:08 pm
by NoMoreQuarantine
NormalDiffusion wrote:What was wrong? Was having a quick look this afternoon...
The way I had done it was comparing "frame size" in atoms vs time/atoms processed. The problem is that that is not how OpenMM works; OpenMM simulates all of the atoms at once applying an integrator in steps. A frame is a snapshot of 1% of those integrator steps. I misunderstood the process and made charts that were incorrect.

I believe the correct way to do it is to simply compare the atom count to the time it takes to process 1 frame (TPF). Which I have done here: https://imgur.com/gallery/nXFM3cV

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 7:22 pm
by Kebast
HFM is showing project 13872 at 727k PPD for my CPU. Normal range is around 160k.
That an HFM issue?

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 7:24 pm
by NoMoreQuarantine
Kebast wrote:HFM is showing project 13872 at 727k PPD for my CPU. Normal range is around 160k.
That an HFM issue?
What's your TPF? I can calculate what your PPD should be.

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 7:38 pm
by Kebast
NoMoreQuarantine wrote:
Kebast wrote:HFM is showing project 13872 at 727k PPD for my CPU. Normal range is around 160k.
That an HFM issue?
What's your TPF? I can calculate what your PPD should be.
ProjectID ProjectRun ProjectClone ProjectGen FrameTime
13872 0 1565 121 00:00:16
13872 0 2370 179 00:00:16
13872 0 3206 137 00:00:16

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 7:49 pm
by NoMoreQuarantine
Kebast wrote:ProjectID ProjectRun ProjectClone ProjectGen FrameTime
13872 0 1565 121 00:00:16
13872 0 2370 179 00:00:16
13872 0 3206 137 00:00:16
That project is 154326 atoms large. I calculate the PPD should be 727378.3 with those TPF numbers for a total credit of 13470. It looks like that lines up with with amount of credits you received.

That's pretty crazy. What CPU are you using?

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 8:16 pm
by Kebast
NoMoreQuarantine wrote:
Kebast wrote:ProjectID ProjectRun ProjectClone ProjectGen FrameTime
13872 0 1565 121 00:00:16
13872 0 2370 179 00:00:16
13872 0 3206 137 00:00:16
That project is 154326 atoms large. I calculate the PPD should be 727378.3 with those TPF numbers for a total credit of 13470. It looks like that lines up with with amount of credits you received.

That's pretty crazy. What CPU are you using?
That's odd for sure. CPU is Ryzen 3800x, 15 threads.

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 8:18 pm
by Kebast
Kebast wrote:
NoMoreQuarantine wrote:
Kebast wrote:ProjectID ProjectRun ProjectClone ProjectGen FrameTime
13872 0 1565 121 00:00:16
13872 0 2370 179 00:00:16
13872 0 3206 137 00:00:16
That project is 154326 atoms large. I calculate the PPD should be 727378.3 with those TPF numbers for a total credit of 13470. It looks like that lines up with with amount of credits you received.

That's pretty crazy. What CPU are you using?
That's odd for sure. CPU is Ryzen 3800x, 15 threads.
Something must be off...it's only taking 28 minutes to complete.

Re: PPD Database - HFM exports needed

Posted: Thu Apr 30, 2020 8:23 pm
by NoMoreQuarantine
I just checked my work history, I also had amazing performance on project 13872. Not as good as yours mind you. With my Ryzen 2600 I would normally get around 55K, but on that one I got 191K.