As i need more desktop space i was looking for a possibility to run 4 Monitors. I just got my answer from nvidia support, that i can achieve running 4 Monitors by adding a 2nd Graphics-Card in non-sli setup.
So my question is now: Will i be able to have the advantage to use 2 x 512 CUDA Cores if i run two Nvidia GTX580 Cards in non-sli Mode ?
I think this question is related to the CUDA Bridge, as it's not nvidia related, how the GPU is used by nebula.
This sounds simple, but if loading 20 instances of nebula server, how the cores will be adressed ? Well as i'm a nerd in such things i don't want to know how ; )
AFAICS since my yesterdays testings that CUDA isn't a real option for me at the moment since i can run at least 4 times more 64 bit nebula instances within my local CPU/DAW, than with nebula set in CUDA mode. CUDA IMO is a nice addition and also keeps CPU load lower, but as it's RAM limited the outsourcing of CPU Power and RAM to a server seems to be a much more powerful and flexible option.
Do you really see a future for using nebula with CUDA in comparison to the new nebula server edition ? Or is CUDA just a additional feature / possibility to reach a bit more GPU power ?
Anyhow, just wanted to say that you guys have done a great work with this server version of nebula.
IMO much more GPU power and lesser RAM limits available. Well, the hardware has its costs but at least the limitations are far higher, than with those 1.5 Gb RAM. Even ATI makes cards with more RAM but no CUDA / TESLA support.
nvidia know their biz in the pro league.
And yes, I know that we "only" make music and no weather reports for the coming week, but uhm so just see it as a written thought I'd really love a more powerful system.