|
|
#1 |
|
Modder
Join Date: Apr 2012
Location: Sweden
Posts: 68
![]() |
Quadro NVS series and folding
Hi,
has anyone got any experience on the Quadro NVS series of cards? Can they run F@H at all? and if so any idea of PPDs? I am interested in the 290, 295 and the 300 models. Cheers
__________________
--- Ferramentis venit et vadit. Scientia manet --- |
|
|
|
|
|
#2 |
|
Was that a warranty I just broke?
Join Date: Oct 2011
Location: Northampton
Posts: 1,301
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I don't know if fah supports that generation of card, as it is one of the first and slowest to have compute capabilities. The ppd/W will be very low if they are supported.
__________________
|
|
|
|
|
|
#3 |
|
Modder
Join Date: Apr 2012
Location: Sweden
Posts: 68
![]() |
Humour me with specifics
![]() I think the newer NVS 300 has some support for CUDA. They are all passively cooled (hence silent) so they sound ideal for a no frills no work interrupting folding
__________________
--- Ferramentis venit et vadit. Scientia manet --- |
|
|
|
|
|
#4 |
|
Was that a warranty I just broke?
Join Date: Oct 2011
Location: Northampton
Posts: 1,301
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I can't give confirmed numbers, but I can do some maths.
A GTX 680 is capable of 3250 GFLOPS, and can generate 20-30K ppd An NVS 300 can generate 45 GFLOPS, so I would estimate a maximum of 280-420 ppd. An NVS 295 generates around 21 GFLOPS ==> 130-195 ppd An NVS 290 generates around 30 GFLOPS ==> 185-280 ppd As there is no quick return bonus (yet) for GPUs, the above figures can be considered reasonably accurate, or perhaps even over generous. When Stanford implement the QRB for GPUs, the difference will be even greater
__________________
|
|
|
|
![]() |
| Thread Tools | |
|
|