Multiple GPU's, Diminishing Returns

Moderators: Site Moderators, FAHC Science Team

Post Reply
madwedge
Posts: 2
Joined: Tue Jun 08, 2021 9:59 am

Multiple GPU's, Diminishing Returns

Post by madwedge »

Hello,

I'm entirely new to Folding@Home so please bear with me. I hope this is the right section to ask this. I have an ETH mining rig which consists of five 30 series GPU's. A 3060Ti, three 3070's and a 3080. I also have a 3080 in my separate gaming PC.

I first started Folding the other day with my gaming PC. I'm seeing over 5 million PPD on the 3080. I then got my five card mining rig Folding too, but I'm seeing very reduced rates. Firstly each GPU usage only seems to be around 50% for some reason. Due to this the PPD of each card isn't that high. For example, compared to the PPD of 5 million on my gaming PC, the 3080 on the mining rig is only about 1.5 million.

I appreciate Folding is completely different to Mining, and that double the GPU power doesn't mean double the points per day. But having my cards configured onto one Motherboard in my rig, is this having a negative effect? Would I be better to Fold with just the 3080, or a couple of the 3070's for example?

As I say I'm completely new to Folding and just trying to understand the mechanics of whats involved.

Many thanks
JimboPalmer
Posts: 2573
Joined: Mon Feb 16, 2009 4:12 am
Location: Greenwood MS USA

Re: Multiple GPU's, Diminishing Returns

Post by JimboPalmer »

Welcome to F@H!

Both the PCIE slots and the CPU have what are called Lanes. (think of them as data pumps, you need data and so you need some lanes to get data)

Usually a Motherboard has one slot with x16 lanes then perhaps one with x8 lanes and several with only x1 lanes (the little x is a buzzword) Sometimes slots slow down as you populate more of them.

Linux tends to use the lanes better than wWndows, but in either case, less than x8 lanes slows moving data on and off your graphics card. You obviously move data on and off at the start and end, but you also move data at every checkpoint.

Back at the CPU it also has a limit to the number of lanes, so in some cases, your motherboard may have faster slots with a better CPU.

I cant look up either your CPU or Motherboard as you did not name names, but you can, on google.

here are my specs:

https://www.tomshardware.com/reviews/am ... n-3-review

My CPU only has 20 lanes

https://www.asus.com/us/Motherboards-Co ... /techspec/

My Mother board has one x16 slot and two x1 slots.
Last edited by JimboPalmer on Tue Jun 08, 2021 3:01 pm, edited 1 time in total.
Tsar of all the Rushers
I tried to remain childlike, all I achieved was childish.
A friend to those who want no friends
aetch
Posts: 447
Joined: Thu Jun 25, 2020 3:04 pm
Location: Between chair and keyboard

Re: Multiple GPU's, Diminishing Returns

Post by aetch »

Can you post the first few hundred lines of your log file so we can ogle at the rest of the specs of your machine :wink:
Instructions -> viewtopic.php?f=61&t=26036

Beyond that Folding is somewhat different to mining, the cpu is constantly moving data to and from the gpu.
The cpu should be multicore, the client dedicates a single cpu thread to each gpu, it'll help to have a thread or two left over for the operating system.
The PCIe interface should be at least x4, x8 is recommended. If you're running Linux I believe you can get away with lower requirements as Linux has less overhead.

The better the base system the better the GPU performs. Although I've learned this through trial and error I haven't been able to pinpoint whether the CPU or memory is more important here, others might have.
Folding Rigs - None (25-Jun-2022)

ImageImage
Joe_H
Site Admin
Posts: 7857
Joined: Tue Apr 21, 2009 4:41 pm
Hardware configuration: Mac Pro 2.8 quad 12 GB smp4
MacBook Pro 2.9 i7 8 GB smp2
Location: W. MA

Re: Multiple GPU's, Diminishing Returns

Post by Joe_H »

There are two main areas of differences between mining and F@H that come into play with multi-GPU setups. Both relate to the motherboard and CPU.

First is the CPU. Folding requires the use of one CPU core/thread per GPU to handle preparing data and instructions to be worked on by the GPUs. Many mining rigs use a low end CPU as it is not that important for the mining calculations, and may not support as many threads as needed for a 5 GPU rig.

Second are the PCIe lanes. Mining loads a calculation onto a GPU and every so often gets back results. F@h is continually going through a cycle of preparing data and instructions, loading it onto the GPU, and then retrieving the results to be used for further calculations. This happens many times a second. So slower or narrower PCIe connections will slow down folding, especially on faster GPUs. So the number of PCIe lanes available through the motherboard and also the CPU will be a consideration.

Connected to PCIe utilization, Linux will run more efficiently than Windows. A high end card will work best with at least a PCIe 3 4x or better connection under Linux, 8x may be required to get the same performance under Windows.
Image

iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
madwedge
Posts: 2
Joined: Tue Jun 08, 2021 9:59 am

Re: Multiple GPU's, Diminishing Returns

Post by madwedge »

Yes, since it's a rig designed for mining Ethereum, it does have a poor CPU, a Pentium Gold G6400. 2 Cores, 4 Threads. I've deselcted the CPU from the Slots in F@H app, but I guess it still plays a part in dishing out Work Units? Also all the five GPU's are on mining risers and using x1 riser cards. I can see how both the CPU and risers can effect the efficiancy. Log attached below for those interested. Many thanks

Code: Select all

*********************** Log Started 2021-06-08T06:40:40Z ***********************
06:40:40:******************************* libFAH ********************************
06:40:40:           Date: Oct 20 2020
06:40:40:           Time: 13:36:55
06:40:40:       Revision: 5ca109d295a6245e2a2f590b3d0085ad5e567aeb
06:40:40:         Branch: master
06:40:40:       Compiler: Visual C++ 2015
06:40:40:        Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Zc:throwingNew /MT
06:40:40:       Platform: win32 10
06:40:40:           Bits: 32
06:40:40:           Mode: Release
06:40:40:****************************** FAHClient ******************************
06:40:40:        Version: 7.6.21
06:40:40:         Author: Joseph Coffland <joseph@cauldrondevelopment.com>
06:40:40:      Copyright: 2020 foldingathome.org
06:40:40:       Homepage: https://foldingathome.org/
06:40:40:           Date: Oct 20 2020
06:40:40:           Time: 13:41:04
06:40:40:       Revision: 6efbf0e138e22d3963e6a291f78dcb9c6422a278
06:40:40:         Branch: master
06:40:40:       Compiler: Visual C++ 2015
06:40:40:        Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Zc:throwingNew /MT
06:40:40:       Platform: win32 10
06:40:40:           Bits: 32
06:40:40:           Mode: Release
06:40:40:         Config: C:\ProgramData\FAHClient\config.xml
06:40:40:******************************** CBang ********************************
06:40:40:           Date: Oct 20 2020
06:40:40:           Time: 11:36:18
06:40:40:       Revision: 7e4ce85225d7eaeb775e87c31740181ca603de60
06:40:40:         Branch: master
06:40:40:       Compiler: Visual C++ 2015
06:40:40:        Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Zc:throwingNew /MT
06:40:40:       Platform: win32 10
06:40:40:           Bits: 32
06:40:40:           Mode: Release
06:40:40:******************************* System ********************************
06:40:40:            CPU: Intel(R) Pentium(R) Gold G6400 CPU @ 4.00GHz
06:40:40:         CPU ID: GenuineIntel Family 6 Model 165 Stepping 3
06:40:40:           CPUs: 4
06:40:40:         Memory: 7.85GiB
06:40:40:    Free Memory: 5.75GiB
06:40:40:        Threads: WINDOWS_THREADS
06:40:40:     OS Version: 6.2
06:40:40:    Has Battery: false
06:40:40:     On Battery: false
06:40:40:     UTC Offset: 1
06:40:40:            PID: 9048
06:40:40:            CWD: C:\ProgramData\FAHClient
06:40:40:  Win32 Service: false
06:40:40:             OS: Windows 10 Enterprise
06:40:40:        OS Arch: AMD64
06:40:40:           GPUs: 6
06:40:40:          GPU 0: Bus:2 Slot:0 Func:0 NVIDIA:8 GA104 [GeForce RTX 3070]
06:40:40:          GPU 1: Bus:1 Slot:0 Func:0 NVIDIA:8 GA104 [GeForce RTX 3070]
06:40:40:          GPU 2: Bus:3 Slot:0 Func:0 NVIDIA:8 GA104 [GeForce RTX 3060 Ti]
06:40:40:          GPU 3: Bus:4 Slot:0 Func:0 NVIDIA:8 GA102 [GeForce RTX 3080]
06:40:40:          GPU 4: Bus:5 Slot:0 Func:0 NVIDIA:8 GA104 [GeForce RTX 3070]
06:40:40:          GPU 5: Bus:0 Slot:2 Func:0 INTEL:1 CML GT1 [UHD Graphics 610]
06:40:40:  CUDA Device 0: Platform:0 Device:0 Bus:4 Slot:0 Compute:8.6 Driver:11.3
06:40:40:  CUDA Device 1: Platform:0 Device:1 Bus:1 Slot:0 Compute:8.6 Driver:11.3
06:40:40:  CUDA Device 2: Platform:0 Device:2 Bus:2 Slot:0 Compute:8.6 Driver:11.3
06:40:40:  CUDA Device 3: Platform:0 Device:3 Bus:5 Slot:0 Compute:8.6 Driver:11.3
06:40:40:  CUDA Device 4: Platform:0 Device:4 Bus:3 Slot:0 Compute:8.6 Driver:11.3
06:40:40:OpenCL Device 0: Platform:0 Device:0 Bus:NA Slot:NA Compute:2.1 Driver:27.20
06:40:40:OpenCL Device 1: Platform:1 Device:0 Bus:4 Slot:0 Compute:3.0 Driver:466.27
06:40:40:OpenCL Device 2: Platform:1 Device:1 Bus:1 Slot:0 Compute:3.0 Driver:466.27
06:40:40:OpenCL Device 3: Platform:1 Device:2 Bus:2 Slot:0 Compute:3.0 Driver:466.27
06:40:40:OpenCL Device 4: Platform:1 Device:3 Bus:5 Slot:0 Compute:3.0 Driver:466.27
06:40:40:OpenCL Device 5: Platform:1 Device:4 Bus:3 Slot:0 Compute:3.0 Driver:466.27
06:40:40:***********************************************************************
06:40:40:<config>
06:40:40:  <!-- Folding Slot Configuration -->
06:40:40:  <cause v='HIGH_PRIORITY'/>
06:40:40:
06:40:40:  <!-- Network -->
06:40:40:  <proxy v=':8080'/>
06:40:40:
06:40:40:  <!-- Slot Control -->
06:40:40:  <power v='full'/>
06:40:40:
06:40:40:  <!-- User Information -->
06:40:40:  <passkey v='*****'/>
06:40:40:  <team v='234980'/>
06:40:40:  <user v='a3mex3zrgx2f'/>
06:40:40:
06:40:40:  <!-- Folding Slots -->
06:40:40:  <slot id='1' type='GPU'>
06:40:40:    <pci-bus v='2'/>
06:40:40:    <pci-slot v='0'/>
06:40:40:  </slot>
06:40:40:  <slot id='2' type='GPU'>
06:40:40:    <pci-bus v='1'/>
06:40:40:    <pci-slot v='0'/>
06:40:40:  </slot>
06:40:40:  <slot id='3' type='GPU'>
06:40:40:    <pci-bus v='3'/>
06:40:40:    <pci-slot v='0'/>
06:40:40:  </slot>
06:40:40:  <slot id='4' type='GPU'>
06:40:40:    <pci-bus v='4'/>
06:40:40:    <pci-slot v='0'/>
06:40:40:  </slot>
06:40:40:  <slot id='5' type='GPU'>
06:40:40:    <pci-bus v='5'/>
06:40:40:    <pci-slot v='0'/>
06:40:40:  </slot>
06:40:40:</config>
06:40:40:Trying to access database...
06:40:40:Successfully acquired database lock
06:40:40:FS01:Initialized folding slot 01: gpu:2:0 GA104 [GeForce RTX 3070]
06:40:40:FS02:Initialized folding slot 02: gpu:1:0 GA104 [GeForce RTX 3070]
06:40:40:FS03:Initialized folding slot 03: gpu:3:0 GA104 [GeForce RTX 3060 Ti]
06:40:40:FS04:Initialized folding slot 04: gpu:4:0 GA102 [GeForce RTX 3080]
06:40:40:FS05:Initialized folding slot 05: gpu:5:0 GA104 [GeForce RTX 3070]
06:40:40:WARNING:FS00:Guessing ambiguous GPU to OpenCL device mapping for 00: gpu:0:2 CML GT1 [UHD Graphics 610].  Consider upgrading your graphics driver or manually setting ``opencl-index`` in this slot's configuration.
06:40:40:WARNING:FS00:Disabling beta GPU slot 00: gpu:0:2 CML GT1 [UHD Graphics 610].  Beta GPUs can be tested for no points by setting ``gpu-beta=true`` in the configuration.
06:40:40:WU02:FS03:Starting
06:40:40:WU02:FS03:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:\ProgramData\FAHClient\cores/cores.foldingathome.org/win/64bit/22-0.0.13/Core_22.fah/FahCore_22.exe -dir 02 -suffix 01 -version 706 -lifeline 9048 -checkpoint 15 -opencl-platform 1 -opencl-device 4 -cuda-device 4 -gpu-vendor nvidia -gpu 4 -gpu-usage 100
06:40:41:WU02:FS03:Started FahCore on PID 9192
06:40:41:WU02:FS03:Core PID:6572
06:40:41:WU02:FS03:FahCore 0x22 started
06:40:41:WU01:FS05:Starting
06:40:41:WU01:FS05:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:\ProgramData\FAHClient\cores/cores.foldingathome.org/win/64bit/22-0.0.13/Core_22.fah/FahCore_22.exe -dir 01 -suffix 01 -version 706 -lifeline 9048 -checkpoint 15 -opencl-platform 1 -opencl-device 3 -cuda-device 3 -gpu-vendor nvidia -gpu 3 -gpu-usage 100
06:40:41:WU01:FS05:Started FahCore on PID 8336
06:40:41:WU01:FS05:Core PID:8440
06:40:41:WU01:FS05:FahCore 0x22 started
06:40:41:WU00:FS01:Starting
06:40:41:WU00:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:\ProgramData\FAHClient\cores/cores.foldingathome.org/win/64bit/22-0.0.13/Core_22.fah/FahCore_22.exe -dir 00 -suffix 01 -version 706 -lifeline 9048 -checkpoint 15 -opencl-platform 1 -opencl-device 2 -cuda-device 2 -gpu-vendor nvidia -gpu 2 -gpu-usage 100
06:40:41:WU00:FS01:Started FahCore on PID 8376
06:40:41:WU00:FS01:Core PID:8304
06:40:41:WU00:FS01:FahCore 0x22 started
06:40:41:WU05:FS02:Starting
06:40:41:WU05:FS02:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:\ProgramData\FAHClient\cores/cores.foldingathome.org/win/64bit/22-0.0.13/Core_22.fah/FahCore_22.exe -dir 05 -suffix 01 -version 706 -lifeline 9048 -checkpoint 15 -opencl-platform 1 -opencl-device 1 -cuda-device 1 -gpu-vendor nvidia -gpu 1 -gpu-usage 100
06:40:41:WU05:FS02:Started FahCore on PID 7240
06:40:42:WU05:FS02:Core PID:7428
06:40:42:WU05:FS02:FahCore 0x22 started
06:40:42:WU01:FS05:0x22:*********************** Log Started 2021-06-08T06:40:41Z ***********************
06:40:42:WU01:FS05:0x22:*************************** Core22 Folding@home Core ***************************
06:40:42:WU01:FS05:0x22:       Core: Core22
06:40:42:WU01:FS05:0x22:       Type: 0x22
06:40:42:WU01:FS05:0x22:    Version: 0.0.13
06:40:42:WU01:FS05:0x22:     Author: Joseph Coffland <joseph@cauldrondevelopment.com>
06:40:42:WU01:FS05:0x22:  Copyright: 2020 foldingathome.org
06:40:42:WU01:FS05:0x22:   Homepage: https://foldingathome.org/
06:40:42:WU01:FS05:0x22:       Date: Sep 19 2020
06:40:42:WU01:FS05:0x22:       Time: 02:35:58
06:40:42:WU01:FS05:0x22:   Revision: 571cf95de6de2c592c7c3ed48fcfb2e33e9ea7d3
06:40:42:WU02:FS03:0x22:*********************** Log Started 2021-06-08T06:40:41Z ***********************
06:40:42:WU01:FS05:0x22:     Branch: core22-0.0.13
06:40:42:WU02:FS03:0x22:*************************** Core22 Folding@home Core ***************************
06:40:42:WU01:FS05:0x22:   Compiler: Visual C++ 2015
06:40:42:WU02:FS03:0x22:       Core: Core22
06:40:42:WU01:FS05:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU02:FS03:0x22:       Type: 0x22
06:40:42:WU01:FS05:0x22:             -DOPENMM_GIT_HASH="\"189320d0\""
06:40:42:WU02:FS03:0x22:    Version: 0.0.13
06:40:42:WU01:FS05:0x22:   Platform: win32 10
06:40:42:WU02:FS03:0x22:     Author: Joseph Coffland <joseph@cauldrondevelopment.com>
06:40:42:WU01:FS05:0x22:       Bits: 64
06:40:42:WU02:FS03:0x22:  Copyright: 2020 foldingathome.org
06:40:42:WU01:FS05:0x22:       Mode: Release
06:40:42:WU02:FS03:0x22:   Homepage: https://foldingathome.org/
06:40:42:WU01:FS05:0x22:Maintainers: John Chodera <john.chodera@choderalab.org> and Peter Eastman
06:40:42:WU02:FS03:0x22:       Date: Sep 19 2020
06:40:42:WU01:FS05:0x22:             <peastman@stanford.edu>
06:40:42:WU02:FS03:0x22:       Time: 02:35:58
06:40:42:WU01:FS05:0x22:       Args: -dir 01 -suffix 01 -version 706 -lifeline 8336 -checkpoint 15
06:40:42:WU02:FS03:0x22:   Revision: 571cf95de6de2c592c7c3ed48fcfb2e33e9ea7d3
06:40:42:WU01:FS05:0x22:             -opencl-platform 1 -opencl-device 3 -cuda-device 3 -gpu-vendor
06:40:42:WU02:FS03:0x22:     Branch: core22-0.0.13
06:40:42:WU01:FS05:0x22:             nvidia -gpu 3 -gpu-usage 100
06:40:42:WU02:FS03:0x22:   Compiler: Visual C++ 2015
06:40:42:WU01:FS05:0x22:************************************ libFAH ************************************
06:40:42:WU04:FS04:Starting
06:40:42:WU02:FS03:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU01:FS05:0x22:       Date: Sep 7 2020
06:40:42:WU02:FS03:0x22:             -DOPENMM_GIT_HASH="\"189320d0\""
06:40:42:WU01:FS05:0x22:       Time: 19:09:56
06:40:42:WU04:FS04:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:\ProgramData\FAHClient\cores/cores.foldingathome.org/win/64bit/22-0.0.13/Core_22.fah/FahCore_22.exe -dir 04 -suffix 01 -version 706 -lifeline 9048 -checkpoint 15 -opencl-platform 1 -opencl-device 0 -cuda-device 0 -gpu-vendor nvidia -gpu 0 -gpu-usage 100
06:40:42:WU02:FS03:0x22:   Platform: win32 10
06:40:42:WU01:FS05:0x22:   Revision: 44301ed97b996b63fe736bb8073f22209cb2b603
06:40:42:WU04:FS04:Started FahCore on PID 8472
06:40:42:WU02:FS03:0x22:       Bits: 64
06:40:42:WU01:FS05:0x22:     Branch: HEAD
06:40:42:WU02:FS03:0x22:       Mode: Release
06:40:42:WU01:FS05:0x22:   Compiler: Visual C++ 2015
06:40:42:WU02:FS03:0x22:Maintainers: John Chodera <john.chodera@choderalab.org> and Peter Eastman
06:40:42:WU01:FS05:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU02:FS03:0x22:             <peastman@stanford.edu>
06:40:42:WU01:FS05:0x22:   Platform: win32 10
06:40:42:WU02:FS03:0x22:       Args: -dir 02 -suffix 01 -version 706 -lifeline 9192 -checkpoint 15
06:40:42:WU01:FS05:0x22:       Bits: 64
06:40:42:WU02:FS03:0x22:             -opencl-platform 1 -opencl-device 4 -cuda-device 4 -gpu-vendor
06:40:42:WU01:FS05:0x22:       Mode: Release
06:40:42:WU02:FS03:0x22:             nvidia -gpu 4 -gpu-usage 100
06:40:42:WU01:FS05:0x22:************************************ CBang *************************************
06:40:42:WU02:FS03:0x22:************************************ libFAH ************************************
06:40:42:WU01:FS05:0x22:       Date: Sep 7 2020
06:40:42:WU02:FS03:0x22:       Date: Sep 7 2020
06:40:42:WU01:FS05:0x22:       Time: 19:08:30
06:40:42:WU02:FS03:0x22:       Time: 19:09:56
06:40:42:WU01:FS05:0x22:   Revision: 33fcfc2b3ed2195a423606a264718e31e6b3903f
06:40:42:WU02:FS03:0x22:   Revision: 44301ed97b996b63fe736bb8073f22209cb2b603
06:40:42:WU01:FS05:0x22:     Branch: HEAD
06:40:42:WU02:FS03:0x22:     Branch: HEAD
06:40:42:WU01:FS05:0x22:   Compiler: Visual C++ 2015
06:40:42:WU02:FS03:0x22:   Compiler: Visual C++ 2015
06:40:42:WU01:FS05:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU02:FS03:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU01:FS05:0x22:   Platform: win32 10
06:40:42:WU02:FS03:0x22:   Platform: win32 10
06:40:42:WU01:FS05:0x22:       Bits: 64
06:40:42:WU02:FS03:0x22:       Bits: 64
06:40:42:WU01:FS05:0x22:       Mode: Release
06:40:42:WU02:FS03:0x22:       Mode: Release
06:40:42:WU01:FS05:0x22:************************************ System ************************************
06:40:42:WU02:FS03:0x22:************************************ CBang *************************************
06:40:42:WU01:FS05:0x22:        CPU: Intel(R) Pentium(R) Gold G6400 CPU @ 4.00GHz
06:40:42:WU02:FS03:0x22:       Date: Sep 7 2020
06:40:42:WU01:FS05:0x22:     CPU ID: GenuineIntel Family 6 Model 165 Stepping 3
06:40:42:WU02:FS03:0x22:       Time: 19:08:30
06:40:42:WU01:FS05:0x22:       CPUs: 4
06:40:42:WU02:FS03:0x22:   Revision: 33fcfc2b3ed2195a423606a264718e31e6b3903f
06:40:42:WU01:FS05:0x22:     Memory: 7.85GiB
06:40:42:WU02:FS03:0x22:     Branch: HEAD
06:40:42:WU01:FS05:0x22:Free Memory: 5.64GiB
06:40:42:WU02:FS03:0x22:   Compiler: Visual C++ 2015
06:40:42:WU01:FS05:0x22:    Threads: WINDOWS_THREADS
06:40:42:WU02:FS03:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU01:FS05:0x22: OS Version: 6.2
06:40:42:WU02:FS03:0x22:   Platform: win32 10
06:40:42:WU01:FS05:0x22:Has Battery: false
06:40:42:WU02:FS03:0x22:       Bits: 64
06:40:42:WU01:FS05:0x22: On Battery: false
06:40:42:WU02:FS03:0x22:       Mode: Release
06:40:42:WU01:FS05:0x22: UTC Offset: 1
06:40:42:WU02:FS03:0x22:************************************ System ************************************
06:40:42:WU01:FS05:0x22:        PID: 8440
06:40:42:WU02:FS03:0x22:        CPU: Intel(R) Pentium(R) Gold G6400 CPU @ 4.00GHz
06:40:42:WU01:FS05:0x22:        CWD: C:\ProgramData\FAHClient\work
06:40:42:WU02:FS03:0x22:     CPU ID: GenuineIntel Family 6 Model 165 Stepping 3
06:40:42:WU01:FS05:0x22:************************************ OpenMM ************************************
06:40:42:WU02:FS03:0x22:       CPUs: 4
06:40:42:WU01:FS05:0x22:   Revision: 189320d0
06:40:42:WU02:FS03:0x22:     Memory: 7.85GiB
06:40:42:WU01:FS05:0x22:********************************************************************************
06:40:42:WU02:FS03:0x22:Free Memory: 5.64GiB
06:40:42:WU01:FS05:0x22:Project: 13446 (Run 8405, Clone 33, Gen 0)
06:40:42:WU02:FS03:0x22:    Threads: WINDOWS_THREADS
06:40:42:WU01:FS05:0x22:Unit: 0x00000000000000000000000000000000
06:40:42:WU02:FS03:0x22: OS Version: 6.2
06:40:42:WU01:FS05:0x22:Digital signatures verified
06:40:42:WU02:FS03:0x22:Has Battery: false
06:40:42:WU01:FS05:0x22:Folding@home GPU Core22 Folding@home Core
06:40:42:WU02:FS03:0x22: On Battery: false
06:40:42:WU00:FS01:0x22:*********************** Log Started 2021-06-08T06:40:41Z ***********************
06:40:42:WU01:FS05:0x22:Version 0.0.13
06:40:42:WU02:FS03:0x22: UTC Offset: 1
06:40:42:WU00:FS01:0x22:*************************** Core22 Folding@home Core ***************************
06:40:42:WU02:FS03:0x22:        PID: 6572
06:40:42:WU00:FS01:0x22:       Core: Core22
06:40:42:WU02:FS03:0x22:        CWD: C:\ProgramData\FAHClient\work
06:40:42:WU00:FS01:0x22:       Type: 0x22
06:40:42:WU02:FS03:0x22:************************************ OpenMM ************************************
06:40:42:WU00:FS01:0x22:    Version: 0.0.13
06:40:42:WU02:FS03:0x22:   Revision: 189320d0
06:40:42:WU00:FS01:0x22:     Author: Joseph Coffland <joseph@cauldrondevelopment.com>
06:40:42:WU02:FS03:0x22:********************************************************************************
06:40:42:WU00:FS01:0x22:  Copyright: 2020 foldingathome.org
06:40:42:WU02:FS03:0x22:Project: 17804 (Run 0, Clone 97, Gen 48)
06:40:42:WU00:FS01:0x22:   Homepage: https://foldingathome.org/
06:40:42:WU02:FS03:0x22:Unit: 0x00000000000000000000000000000000
06:40:42:WU00:FS01:0x22:       Date: Sep 19 2020
06:40:42:WU02:FS03:0x22:Digital signatures verified
06:40:42:WU04:FS04:Core PID:8604
06:40:42:WU00:FS01:0x22:       Time: 02:35:58
06:40:42:WU02:FS03:0x22:Folding@home GPU Core22 Folding@home Core
06:40:42:WU04:FS04:FahCore 0x22 started
06:40:42:WU00:FS01:0x22:   Revision: 571cf95de6de2c592c7c3ed48fcfb2e33e9ea7d3
06:40:42:WU02:FS03:0x22:Version 0.0.13
06:40:42:WU00:FS01:0x22:     Branch: core22-0.0.13
06:40:42:WU00:FS01:0x22:   Compiler: Visual C++ 2015
06:40:42:WU00:FS01:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU00:FS01:0x22:             -DOPENMM_GIT_HASH="\"189320d0\""
06:40:42:WU00:FS01:0x22:   Platform: win32 10
06:40:42:WU00:FS01:0x22:       Bits: 64
06:40:42:WU00:FS01:0x22:       Mode: Release
06:40:42:WU00:FS01:0x22:Maintainers: John Chodera <john.chodera@choderalab.org> and Peter Eastman
06:40:42:WU00:FS01:0x22:             <peastman@stanford.edu>
06:40:42:WU00:FS01:0x22:       Args: -dir 00 -suffix 01 -version 706 -lifeline 8376 -checkpoint 15
06:40:42:WU00:FS01:0x22:             -opencl-platform 1 -opencl-device 2 -cuda-device 2 -gpu-vendor
06:40:42:WU00:FS01:0x22:             nvidia -gpu 2 -gpu-usage 100
06:40:42:WU00:FS01:0x22:************************************ libFAH ************************************
06:40:42:WU00:FS01:0x22:       Date: Sep 7 2020
06:40:42:WU00:FS01:0x22:       Time: 19:09:56
06:40:42:WU00:FS01:0x22:   Revision: 44301ed97b996b63fe736bb8073f22209cb2b603
06:40:42:WU00:FS01:0x22:     Branch: HEAD
06:40:42:WU00:FS01:0x22:   Compiler: Visual C++ 2015
06:40:42:WU00:FS01:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU00:FS01:0x22:   Platform: win32 10
06:40:42:WU00:FS01:0x22:       Bits: 64
06:40:42:WU00:FS01:0x22:       Mode: Release
06:40:42:WU00:FS01:0x22:************************************ CBang *************************************
06:40:42:WU00:FS01:0x22:       Date: Sep 7 2020
06:40:42:WU00:FS01:0x22:       Time: 19:08:30
06:40:42:WU00:FS01:0x22:   Revision: 33fcfc2b3ed2195a423606a264718e31e6b3903f
06:40:42:WU00:FS01:0x22:     Branch: HEAD
06:40:42:WU00:FS01:0x22:   Compiler: Visual C++ 2015
06:40:42:WU00:FS01:0x22:    Options: /TP /nologo /EHa /wd4297 /wd4103 /O2 /Ob3 /Zc:throwingNew /MT
06:40:42:WU00:FS01:0x22:   Platform: win32 10
06:40:42:WU00:FS01:0x22:       Bits: 64
06:40:42:WU00:FS01:0x22:       Mode: Release
06:40:42:WU00:FS01:0x22:************************************ System ************************************
06:40:42:WU00:FS01:0x22:        CPU: Intel(R) Pentium(R) Gold G6400 CPU @ 4.00GHz
06:40:42:WU00:FS01:0x22:     CPU ID: GenuineIntel Family 6 Model 165 Stepping 3
06:40:42:WU00:FS01:0x22:       CPUs: 4
06:40:42:WU00:FS01:0x22:     Memory: 7.85GiB
06:40:42:WU00:FS01:0x22:Free Memory: 5.63GiB
06:40:42:WU00:FS01:0x22:    Threads: WINDOWS_THREADS
06:40:42:WU00:FS01:0x22: OS Version: 6.2
06:40:42:WU00:FS01:0x22:Has Battery: false
06:40:42:WU00:FS01:0x22: On Battery: false
06:40:42:WU00:FS01:0x22: UTC Offset: 1
06:40:42:WU00:FS01:0x22:        PID: 8304
06:40:42:WU00:FS01:0x22:        CWD: C:\ProgramData\FAHClient\work
iero
Posts: 177
Joined: Tue Feb 09, 2021 10:40 am

Re: Multiple GPU's, Diminishing Returns

Post by iero »

Welcome to F@H!

As others stated before my, that CPU could run at best 4 GPUS, but don't even think of moving the mouse after F@h starts chugging along.
As for the risers, the x1 bandwidth is faaaar from enough for F@h. Given that your cards are new and powerful, PCIe 3.0 x8 is a must. That 3080 could potentially benefit from a x16 connection.
Image
GPU only
RTX 3060 12GB Gigabyte Gaming OC [currently mining]
Folding since 14/02/2021
MeeLee
Posts: 1375
Joined: Tue Feb 19, 2019 10:16 pm

Re: Multiple GPU's, Diminishing Returns

Post by MeeLee »

Yes, that CPU at best can support 4 GPUs.
That is under Linux.
Also, you'd have to use at least PCIE 4x lanes on each GPU in Linux (3060, 2080Ti).
3070 and 3080 may need more lanes.
Put your fastest GPU in the main slot, and the slower GPUs in the slower (x4) slots (they may be full size, but operating at x4 speeds).
Usually you'll find most motherboards support no more than 2 GPUs nowadays, and Intel motherboards support 3 GPUs max.
One step up, and you'll be able to run server hardware, which can be extremely costly.
bruce
Posts: 20910
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: Multiple GPU's, Diminishing Returns

Post by bruce »

I think the critical information here is that 1x risers are fine for mining but certainly will choke FAH performance on those more powerful GPUs.
Post Reply