is BOINC only CPU related

Message boards : Number crunching : is BOINC only CPU related

To post messages, you must log in.

AuthorMessage
mevanchik

Send message
Joined: 11 Nov 15
Posts: 6
Credit: 1,111,473
RAC: 0
Message 79125 - Posted: 28 Nov 2015, 17:10:04 UTC

I have my old bitcoin machine and its doing terribly compared to my dual xeon cpu server.
is there a setting i have to change, or different team i have to join?
ID: 79125 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Timo
Avatar

Send message
Joined: 9 Jan 12
Posts: 185
Credit: 45,649,459
RAC: 0
Message 79126 - Posted: 29 Nov 2015, 0:17:34 UTC
Last modified: 29 Nov 2015, 0:18:13 UTC

BOINC itself has projects that run on both CPU and on GPUs.

However, this project (Rosetta@Home) performs calculations that are not easily done on a GPU and thus it only uses the CPU and wont use your GPU for anything.

With that said, some other BOINC projects can make use of the GPU to massively speed up computations. If you're specifically interested in biology/medical research/proteins, I suggest looking at Poem@Home or Folding@Home (note that Folding@Home is seperate from the BOINC platform.) Both of these projects will make good use of your GPUs to accelerate computations.

With that said, Rosetta is a fantastic project when it comes to improving our understanding of protein structures, folding, docking, etc. and I'd encourage you to keep your Xeon server contributing to Rosetta@Home.
ID: 79126 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
sgaboinc

Send message
Joined: 2 Apr 14
Posts: 282
Credit: 208,966
RAC: 0
Message 79131 - Posted: 29 Nov 2015, 19:36:24 UTC - in response to Message 79126.  

i read folding@home uses molecular dynamics algorithms which may be somewhat different from the algorithms rosetta@home uses, the different modelling approach may in part make the use of gpu limited (but only the researchers could tell)

https://en.wikipedia.org/wiki/Molecular_dynamics
https://en.wikipedia.org/wiki/Molecular_modeling_on_GPU
https://en.wikipedia.org/wiki/GROMACS
ID: 79131 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile dcdc

Send message
Joined: 3 Nov 05
Posts: 1832
Credit: 119,664,803
RAC: 11,191
Message 79134 - Posted: 29 Nov 2015, 21:45:01 UTC

As Timo and sgaboinc have said, rosetta@home can't use your GPU(s), but your 2xXeon machine is doing a great job.

The team you are on makes no difference to any settings - it's just for fun.

The projects you choose within BOINC are a different matter - rosetta@home is one of many (but the best in terms of research outputs and potential in my opinion!).

You could upgrade the celeron CPU to something with more cores - you'd have to check what your motherboard can take, but you could go up to an i7-3770K. An i5 would be cheaper and not much slower.

D
ID: 79134 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
mevanchik

Send message
Joined: 11 Nov 15
Posts: 6
Credit: 1,111,473
RAC: 0
Message 79136 - Posted: 30 Nov 2015, 3:49:04 UTC

thanks guys , i did end up googling and found it.

so i use my xeons here, and my quad saphire's there.

thanks!
ID: 79136 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : is BOINC only CPU related



©2024 University of Washington
https://www.bakerlab.org