when u catch 'em using F@H as a benchmark

Moderators: Site Moderators, FAHC Science Team

Post Reply
Knish
Posts: 232
Joined: Tue Mar 17, 2020 5:20 am

when u catch 'em using F@H as a benchmark

Post by Knish »

u get that feeling of doubt they hit the finish button and waited... cuz they're in the business of other stuff. Yet if u zoom in u can make out the PRGC and so u check... and nothing shows in the WU app site but their powerful GPU should've completed it long ago, or u see usernames that completed it that are definitely not who you're watching. Tech channels, c'mon bruh, u should know better.
BobWilliams757
Posts: 493
Joined: Fri Apr 03, 2020 2:22 pm
Hardware configuration: ASRock X370M PRO4
Ryzen 2400G APU
16 GB DDR4-3200
MSI GTX 1660 Super Gaming X

Re: when u catch 'em using F@H as a benchmark

Post by BobWilliams757 »

Maybe the tech channel couldn't get a stable enough system to fold the WU? :) With the influx of Youtube and everyone thinking they are an expert on everything.... it could happen.
Fold them if you get them!
gunnarre
Posts: 567
Joined: Sun May 24, 2020 7:23 pm
Location: Norway

Re: when u catch 'em using F@H as a benchmark

Post by gunnarre »

PPD (points per day) is a fairly good number for checking if your system is a good folding rig, but PPD is a terrible benchmark for A-B comparisons of systems and configurations, because the PPD is very dependent on how quickly you return the WU and the Quick Return Bonus, as well as variance between projects. If you run the same WU twice, the second time will give you a lower PPD estimate because you're later in the alloted time for returning it.

If you want to benchmark with Folding@Home, you should look at the TPF (time per frame) number on the same project, and ideally you should fold the exact same WU on several computers if you want a fair benchmark. And as you say they should fold the WU to completion and upload the results on at least one computer before the Timeout time, because anything else ties up the science by making the project wait.
Image
Online: GTX 1660 Super, GTX 1080, GTX 1050 Ti 4G OC, RX580 + occasional CPU folding in the cold.
Offline: Radeon HD 7770, GTX 960, GTX 950
Neil-B
Posts: 2027
Joined: Sun Mar 22, 2020 5:52 pm
Hardware configuration: 1: 2x Xeon E5-2697v3@2.60GHz, 512GB DDR4 LRDIMM, SSD Raid, Win10 Ent 20H2, Quadro K420 1GB, FAH 7.6.21
2: Xeon E3-1505Mv5@2.80GHz, 32GB DDR4, NVME, Win10 Pro 20H2, Quadro M1000M 2GB, FAH 7.6.21 (actually have two of these)
3: i7-960@3.20GHz, 12GB DDR3, SSD, Win10 Pro 20H2, GTX 750Ti 2GB, GTX 1080Ti 11GB, FAH 7.6.21
Location: UK

Re: when u catch 'em using F@H as a benchmark

Post by Neil-B »

Actually if you want to benchmark with FaH then far better to take a copy of the working folder for a WU you are completing and run the Fahcore from command line "offline" ... You can use the same wu time and time again on all kit and have a direct comparison of performance ... also useful if one feels the need to tweak settings and wants to ensure FaH is still stable without risking stuffing up the science ... no points as such for running wu this way (never actually connects/sends the results simply runs/loads the kit) but does far less damage.
2x Xeon E5-2697v3, 512GB DDR4 LRDIMM, SSD Raid, W10-Ent, Quadro K420
Xeon E3-1505Mv5, 32GB DDR4, NVME, W10-Pro, Quadro M1000M
i7-960, 12GB DDR3, SSD, W10-Pro, GTX1080Ti
i9-10850K, 64GB DDR4, NVME, W11-Pro, RTX3070

(Green/Bold = Active)
Post Reply