New GPU3 client released

S

SazanEyes

[H]ard|DCer of the Month - January 2011
Joined
Jan 14, 2009
Messages
1,250
From Vijay's blog:

New GPU3 (version 6.41) client released

The latest GPU3 client is now on our high performance client download page. This new client improves GPU identification and has several other minor improvements.

For those who are unfamiliar with the other FAH clients: there are several types of high performance clients. SMP refers to multi-core GPUs. GPU3 refers to the latest generation Folding@home GPU client. Currently, GPU3 is only available for NVIDIA GPUs, but we are working on a GPU3 release for ATI. For NVIDIA GPUs, we recommend GPU3. For ATI, we suggest GPU2 until GPU3 has passed through its QA testing (going on now). Finally, the v7 Folding@home client (currently in alpha testing) has these functions (SMP and GPU) built in and will not require a special client.
Click to expand...
 
So this means they can send larger WU to cards that can handel them?
 
gkanakis said:
Safe to say that I need to stay with GPU2 with my GTX200 series cards?
Click to expand...

GPU3 is only for Fermi Chips (400 or higher models). I installed the 6.40 console client earlier, didn't notice any change so far.
 
TheMTZ said:
GPU3 is only for Fermi Chips (400 or higher models). I installed the 6.40 console client earlier, didn't notice any change so far.
Click to expand...
GPU3 works with most CUDA-compatible GPUs, including G8x, G9x, GT2xx, and GF1xx. However, GT2xx GPUs perform better with GPU2.
 
Do i have to do anything if im using GPU tracker? or will it automagically download it?
 
For the record.... GPU3 is for ATI cards too. It just doesn't do any new work.

BUT it does properly recognise them.

You'd need to uninstall the gpu3 client... as this is the part that downloads each core... but no idea if tracker would have updated the d/l part.
 
I really wonder how ATI cards will perform once we have proper support for them.
 
ok lets stop the confusion.. the gpu3 client works for everything. it can run gpu2 or gpu3 WU's depending on the flags you run. there are 3 set's of WU's. Fermi GPU3, non fermi GPU3, and GPU2.
 
Can someone inform me on the proper flag to use with this new client and G92 cards in order to receive GPU3 WUs (P1117x)? I tried the -forcegpu nvidia_g80_1.0 flag and all I'm receiving are GPU2 WUs. Should I use the '1.1' or another iteration?

TIA
 
Don't use any -forcegpu flags. The client should automatically detect the highest CUDA version supported by the cards and download units accordingly.
 
Zero82z said:
Don't use any -forcegpu flags. The client should automatically detect the highest CUDA version supported by the cards and download units accordingly.
Click to expand...
OK, I'll remove the force flag, that should simply things.
 
You must log in or register to reply here.
Back
Top