Page 15 of 17

Re: PPD Database - HFM exports needed

Posted: Mon May 18, 2020 6:37 pm
by Juggy
NoMoreQuarantine wrote:
Juggy wrote:In the greater scheme of things a KO is just a regular 2060 and may or may not get the Turing 104 chip right so why is it classed differently to a standard 2060?

Or, are you treating that as extenuating circumstances in which case it may be a good idea to identify which ones have the TU104 and which have the TU106 chip.

My point and suggestion was to split the cards out even more granualary so people can make a decision to purchase based on an actual manufacturer/model. This may very well be too much work and maybe not worth the hassle anyway.

Just a thought
EVGA has stated that all 2060 KOs will come with the TU104. They are the only manufacturer who has done this as far as I am aware. I can confirm that mine is a TU104 according to GPU-Z. That said, it may make sense to relabel it in the database as "2060 TU104" as they do on Techpowerup and leave TU106s labeled as just 2060s.
Oh, didn't realise they had stated that. The last video I watched seemed to indicate that you may or may not get the TU104 in the KO.

Re: PPD Database - HFM exports needed

Posted: Mon May 18, 2020 6:42 pm
by NoMoreQuarantine
Juggy wrote:Oh, didn't realise they had stated that. The last video I watched seemed to indicate that you may or may not get the TU104 in the KO.
A lot of people got that impression (myself included), but Gamers Nexus clarified EVGAs statement in a followup video.

Re: PPD Database - HFM exports needed

Posted: Mon May 18, 2020 6:46 pm
by NoMoreQuarantine
NormalDiffusion wrote:Just to add a little bit more to the confusion. I got 2 more Radeon vii, so 3 in total. What I can see is just crazy! One was performing with 1/3 of the ppd. Moving her to another pc brought her to the normal value (I thought she was doa). The pc where she got 1/3 of the ppd is a e5-160 v4, so not really slow. On a gen2 16x Pci-e the radeon vii is loosing 30% of her ppd...
on the other side, a 2070super on the e5-1650 v4 has normal ppd. As well as on a gen2 16x Pci-e (xeon x5670 to be precise...).
Wondering why I was not writing and submitting data the last days? Well, I'm in hell...
And trying Linux made it even worse...
That sounds horrible :(

Let us know if you discover the reason for the drop in performance. That seems like something we would want to be aware of. Do you think it could have just been an unusual project?

Re: PPD Database - HFM exports needed

Posted: Mon May 18, 2020 7:49 pm
by NormalDiffusion
NoMoreQuarantine wrote: That sounds horrible :(

Let us know if you discover the reason for the drop in performance. That seems like something we would want to be aware of. Do you think it could have just been an unusual project?
I don't think it's an unusual project. I already had a couple of clones of this project running on the same rvii without problem on a different pc (1.2Mppd vs 0.4Mppd on the e5-1650 v4). Testing with fahbench (the modified version with core 22) showed a 50% difference to another pc.
I will check the Pci-e slot and clean it, as well as the vrms and ram temperature on the rvii.

Also, I don't get why the rvii is so castrated with a Pci-e Gen 2 when on the same machine a 2070super is running full speed with 30-35% bus usage for fah (under win 10).

EDIT: typos

Re: PPD Database - HFM exports needed

Posted: Tue May 19, 2020 12:20 am
by BobWilliams757
NoMoreQuarantine wrote:
lafrad wrote:Pretty sure the general idea for this was... to get a general idea. take the data from a spectrum of situations, and give an idea where you expect to be. Things as subtle as the brand of case fan will adjust the PPD of an air-cooled GFX card..... why worry about that point?
Yep, I am pretty happy with the current data. I feel it gives a good idea of how someone could expect a processor to perform, and supplies the raw data that can be used to perform analysis across the different projects and atom counts.

What it doesn't tell us is how energy efficient or cost effective the individual hardware is. Which are separate practical issues many of us are interested in.
Agreed, but having all the data for the specifics might be a nightmare to keep up with.

And realistically, we will probably have people upping or lowering power limits, searching for sweet spots, etc.... even if they had the exact same card.

I think it should be a good enough gauge for anyone hardware shopping to give them a good starting point on the suitability for folding. And if someone wants specifics, a quick search of this thread would probably help them find users of that hardware quickly, and get that info.

Re: PPD Database - HFM exports needed

Posted: Tue May 19, 2020 1:26 am
by lafrad
You could move to the geometric mean: https://en.wikipedia.org/wiki/Geometric_mean

This would normalize things a bit more and should reduce the impact of "over-reporting"....

OR, average each users data first, then average the averages (instead of all the datapoints)...

Lots of ways to process the data to account for this stuff....

Re: PPD Database - HFM exports needed

Posted: Fri May 22, 2020 10:58 am
by NormalDiffusion
NoMoreQuarantine wrote:
Let us know if you discover the reason for the drop in performance. That seems like something we would want to be aware of. Do you think it could have just been an unusual project?
Problem solved on the xeon 1650 v4. I had to disable the c-states. She's going strong now!

I might have discovered something else, just not 100% sure about it right now. To check I had to order a new cpu, so as soon as I can test I'll give a follow up.
Spoiler alert:
It has to do with the GHz myth

Re: PPD Database - HFM exports needed

Posted: Sat May 23, 2020 2:52 pm
by BobWilliams757
Just a note to anyone that is submitting data. Keep an eye on your FAH logs vs HFM. I've got one particular WU that consistently reports weird on HFM vs actual FAH logs and awarded points. In this case I think the WU might actually have a longer timeout than it displays, but I'm not sure. And it's one of those WU's that performs gives some GPU's issues as well, and uses more CPU than most of them do.

Being that I keep getting those WU's, I'd have to edit the HFM export to really find the better numbers for my APU.

Re: PPD Database - HFM exports needed

Posted: Tue May 26, 2020 8:40 pm
by foldinghomealone
The database is currently under maintanance.
There is an issue that that WU name and path don't match.
I don't know what messed this up, probably it was me.

The PPD and HW stats seem to be ok, though.

Currently I don't have time to set up everything again.
There are too many uploaded single files I have to process manually.

Re: PPD Database - HFM exports needed

Posted: Tue May 26, 2020 9:20 pm
by Kebast
foldinghomealone wrote:The database is currently under maintanance.
There is an issue that that WU name and path don't match.
I don't know what messed this up, probably it was me.

The PPD and HW stats seem to be ok, though.

Currently I don't have time to set up everything again.
There are too many uploaded single files I have to process manually.
If it helps, I could just upload my entire file again and you could replace everything I've given you.

Re: PPD Database - HFM exports needed

Posted: Tue May 26, 2020 9:34 pm
by markdotgooley
My RTX 2060 KO is a TU 106 - 200A, whatever that is. EVGA. I also have a TU 106 in a regular RTX 2060, also EVGA. Beats me.

Re: PPD Database - HFM exports needed

Posted: Tue May 26, 2020 11:59 pm
by NoMoreQuarantine
markdotgooley wrote:My RTX 2060 KO is a TU 106 - 200A, whatever that is. EVGA. I also have a TU 106 in a regular RTX 2060, also EVGA. Beats me.
Is that what Folding@home is telling you or GPU-Z?

Re: PPD Database - HFM exports needed

Posted: Wed May 27, 2020 4:00 am
by NoMoreQuarantine
foldinghomealone wrote:The database is currently under maintanance.
There is an issue that that WU name and path don't match.
I don't know what messed this up, probably it was me.

The PPD and HW stats seem to be ok, though.

Currently I don't have time to set up everything again.
There are too many uploaded single files I have to process manually.
I have a backup I made on the 18th. It looks like the name and path lined up then.

https://docs.google.com/spreadsheets/d/ ... sp=sharing

Re: PPD Database - HFM exports needed

Posted: Wed May 27, 2020 6:45 am
by NormalDiffusion
NoMoreQuarantine wrote: I have a backup I made on the 18th. It looks like the name and path lined up then.

https://docs.google.com/spreadsheets/d/ ... sp=sharing
It's also showing the problem. Check the 16425 projects, you can see it (reported in path as 13xxx).

Re: PPD Database - HFM exports needed

Posted: Wed May 27, 2020 11:32 am
by skydivingcatfan
markdotgooley wrote:My RTX 2060 KO is a TU 106 - 200A, whatever that is. EVGA. I also have a TU 106 in a regular RTX 2060, also EVGA. Beats me.
That is interesting. The opposite of what was reported on GamersNexus and the database on TechPowerup.

How did you determine that you had a TU106?