Fair Play, SC

Vester

Well-Known Member
USA team member
I have been running distributed computing projects since January 2002 starting with United Devices, then Think, then World Community Grid, Rosetta@Home, Folding@home, GpuGrid, Poem and some I may have forgotten. Until a few months ago, I had run Folding@home on two AMD video cards. I became frustrated with poor stats in comparison to the nVidia cards and looked for a project more suited to double precision performance. I am running PrimeGrid double precision (FP64) work units, GFN-22. Since I have two AMD Radeon 7990 video cards, this chart explains why I run GFN-22.

In February (2019), the month that I celebrated my 76th birthday, I bought a rack in order to build a better rig. First photo. The current configuration is running two AMD Radeon HD 7990 video cards. The AMD R9 390 is idle until I win an auction on eBay for a second power supply. I am not running any work units on the six core AMD Ryzen 5 2600 since I have four AMD GPUs running. Recent photo. I prefer access to the business end, not the front. Excuse the wire management, it is a work in progress. The motherboard is a Biostar TM350-BTC Pro. Nothing is overclocked.

I recommend a rack such as mine for accessibility and cooling. It is only 13 inches tall, 13 inches front to back, and 25 1/4 inches wide and is lighter than a case.

I am running Windows 10 Pro because I was unable to get it to crunch Folding@home using Ubuntu and some other distros. I am going to try it again soon with BOINC.

If you ask me any questions, you may get more than a yes or no! :-D
 
Last edited:

Nick Name

Administrator
USA team member
Welcome to the forum, glad to have you aboard! That's a nice setup. Those cards will also do very well on MilkyWay, all their work is FP64. Their work units are much shorter than those PrimeGrid jobs you're currently running, I'm not sure what the change in credit would be. They'll probably also do well on Einstein. Nvidia has nerfed FP64 on their consumer cards for a long time, and OpenCL performance is also much better on AMD. I mostly run Nvidia though because of GPUGrid which right now only runs on Nvidia.
 

Vester

Well-Known Member
USA team member
Thanks for the information. I will try MilkyWay. We all should use our computer assets to their best advantage. Since I am short on monetary assets due to a wife filter, I go to AMD.
 

Nick Name

Administrator
USA team member
I hear you, actually AMD gives a lot more bang for the buck on most projects than Nvidia. If you look at the server status page on Einstein, you will see some shocking numbers comparing AMD output versus Nvidia.

Code:
GPU productivity (last 7 days) 
Hosts w/ ATI/AMD GPU    3,803
Hosts w/ NVIDIA GPU    9,663
Hosts w/ INTEL HD GPU    6,797
Validating hosts w/ ATI/AMD GPU    2,773
Validating hosts w/ NVIDIA GPU    5,138
Validating hosts w/ INTEL GPU    907
Avg daily valid results (ATI/AMD)    50,929
Avg daily valid results (NVIDIA)    41,583
Avg daily valid results (INTEL)    8,572
Avg daily credit (ATI/AMD)    176,463,930
Avg daily credit (NVIDIA)    144,078,400
Avg daily credit (INTEL)    535,759
 
Last edited:

supdood

Well-Known Member
USA team member
Vester--great to have you on the team! It looks like your rig is putting up some great results over at Milky Way!
:USA:
 

Vester

Well-Known Member
USA team member
Vester--great to have you on the team! It looks like your rig is putting up some great results over at Milky Way!
:USA:
It is down for a few days until I replace the motherboard or BIOS chip. Both are on order. You can understand: I was trying to get more output!
 

Vester

Well-Known Member
USA team member
My computer was down more than twenty days while I decided to get an Asus B250 Mining Expert motherboard and an Intel i5 7600 processor. I am using the same Corsair ValueSelect DDR4 2666 RAM, but it runs at DDR4 2400 speed. The computer is identified in the stats as having five Radeon R9 390 video cards. That is wrong. It has one R9 390 and two Radeon HD 7990 (dual GPU) video cards. The RAC is still increasing and is currently 648,245.
 

supdood

Well-Known Member
USA team member
That's great news! Glad to have you back at full strength and crunching for the team.
:USA:
 
Top