Folding Forum

Project 13400 is giving wrong entries in HFM (at least 13400, 4, 8, 0). In my database it appears with a PPD of 229.889,1 and credit of 120.000,0. Actually I got 338.000 credits for the WU (and PPD was around 6xx.xxx, which is in line with a processing time of a little bit more than 12 hours).
Maybe you could check if you see the same behavior and we could think about dropping it from the stats, as it is clearly wrong. Should I be the only one concerned, I'll manually remove it from the dataset before uploading.

HFM is not necessarily at fault here. It checks the list of points per each unit relatively regularly but if you're running Beta units those points allocations can and will change. 13400 changed points up and then back down again within a very brief period. Also, the points indicated by FAH could have been wrong too.

13400-01 are both a little odd. This is not unusual with Beta units.

HaloJones wrote:HFM is not necessarily at fault here. It checks the list of points per each unit relatively regularly but if you're running Beta units those points allocations can and will change. 13400 changed points up and then back down again within a very brief period. Also, the points indicated by FAH could have been wrong too.

13400-01 are both a little odd. This is not unusual with Beta units.

They are not beta WU's, but have been pushed to Advanced. I have only the "Advanced" client-type activated, not beta
If I check the WU, the credit reported by the server match what I got in FAH

Juggy wrote:

foldinghomealone wrote:Thanks to NoMoreQuarantine and Juggy
New updates for:
RTX 2080 Super
RX 590
Intel i9-9900FK
Ryzen 5 2600

My mistake, I labelled the slot incorrectly. It should be i9-9900KF @12t and replace the existing dataset you have

Apologies.

No worries, updated

NormalDiffusion wrote:Project 13400 is giving wrong entries in HFM (at least 13400, 4, 8, 0). In my database it appears with a PPD of 229.889,1 and credit of 120.000,0. Actually I got 338.000 credits for the WU (and PPD was around 6xx.xxx, which is in line with a processing time of a little bit more than 12 hours).
Maybe you could check if you see the same behavior and we could think about dropping it from the stats, as it is clearly wrong. Should I be the only one concerned, I'll manually remove it from the dataset before uploading.

There is no 13400 in the database, did I miss to update one upload?

it was pushed to Advanced to early. It shouldn't be being sent out I think.

foldinghomealone wrote:There is no 13400 in the database, did I miss to update one upload?

It's not in my export, got it yesterday evening
I was taking about my HFM.net database!

HaloJones wrote:it was pushed to Advanced to early. It shouldn't be being sent out I think.

From what I saw, it's doing ok on nVidia cards under Linux. But the Radeon VII under Win 10... Taking ages to complete while doing more or less nothing

There are currently two separate CPU entries "R5 2600" and "Ryzen 5 2600" that should be combined.

NoMoreQuarantine wrote:There are currently two separate CPU entries "R5 2600" and "Ryzen 5 2600" that should be combined.

Thanks, done

New updates for
i7 8700k @12t
i7-2600K
i5-8400
Xeon E5649 @18t
RX Vega 64
RX 580

Thanks HenrikJolsen and AOD_N3URAL

I uploaded some WU's from my Vega 11 while I still have the chance to take the bottom spot!

It looks like the database is getting plenty of scoop. And despite all the talk of some cards liking certain size WU atom counts, it appears that for the most part the ranking stays intact on the project page results.


Whoever submitted the R3 1200 @3t has that thing cooking! I have to wonder if that is stock clocks and that PPD for that processor, but either way it seems solid.

And I'm glad a 1650 Super made a showing, as I'm thinking about one myself.

BobWilliams757 wrote:And I'm glad a 1650 Super made a showing, as I'm thinking about one myself.

That's me. I'll try to get more stats for you, but she's not picking up much WUs...

The R3 1200 @3t is me.

It is running all cores at 3.7Ghz instead on standard 3.1Ghz. The memory is running DDR4 3200.

It's going to be tough to create such a PPD list.
With core 21 we had pretty linear results.
Core 22 can vary wildly depending on the project you get assigned.
For instance, I have been able to run a 2080Ti at 4.4M PPD in Linux on core 22, but in other projects it maxed out at only 3.6M PPD.
Core 21 is a much better benchmarking or reference tool, as most of the data is established.
Not until core 22 projects give WUs that are made for the GPU (no small atom count for larger GPUs like the RTX series of GPUs), can you make an accurate and viable PPD chart.

Aside from differences in cooling solutions (blower type vs open 1 fan, vs open 2 fan , vs open 3 fan), there's operating system (Win vs Linux vs VM), slot speed (pcie 2.0 vs 3.0 vs 4.0, @ x4, x8 or x16), overclocked vs stock, with or without power capped; and most limiting factor of all temps (closed case vs open bench, ambient temps), that can all affect performance by more than 10%!

With newer models being introduced in the next month the list would be extremely extensive, and trusting data from people that could say whatever they want (bump a few extra PPDs just for the fun of it).