OTOY Forums

It'll be a while before I purchase any more video cards after what my current system has cost me, it already cost me double the amount I normally pay to build myself a new system. To be honest, I can't imagine I'd need more than the 1GB I have access to right now, but even so, the option can't hurt if it was possible. Better to be there, than not there.

It'd be a heck of a lot cheaper than buying another card

pumeco wrote:Thanks again, Marcus, much appreciated.

Here's how my CUDA device setup looks now, so this is after removing the Octane 3.2 build, then installing 3.0.

It's noticeably faster now, but what I'm curious about is what it says at the top. It says the CUDA driver version is 3.20, but the CUDA Runtime version is 3.00. So does this mean I'm getting the current maximum performance for now, or should I downgrade the driver to match the Runtime?

No performance wouldn't change, but you potentially re-introduce bugs that have been fixed with newer driver versions. You should be able to use newer CUDA drivers with older CUDA runtime libraries, but not the other way round.

As for the possibly big secret positive side-effect with CUDA 4 ...

All I can say is that if it turns out to be the ability to share the graphics RAM between cards, that would be fantastic. This CUDA stuff is all pretty new to me, but even from my beginner point of view, I can think of no better icing on the cake as being able to pool the RAM from multiple cards and make the total available to Octane.

Yeah I know, I live in a fantasy land

As discussed several times before: That won't happen in the near future. And on the longer term the availability of graphics cards with more VRAM will become better, which removes any need to accept serious performance hits (I'm talking about things being 10-100 times slower) for larger memory.

Sorry, but that's how things are. Cheers,
Marcus

Thanks for the explanation, I'll just leave it as it's set-up right now then.

Regards the RAM, I totally understand that decision. I wasn't aware of the real implications of it. But yup, I'd just add another card with more RAM if ever I needed it. At least the power of the cores of each card added will still get used, so there's that as another added benefit to an additional card as well. Still, I think I'll have to wait until I see how instancing is implemented before I decide if I need more ram.

Things like polygonal grass and hair for example; if it can be instanced without hitting the graphics RAM, I'd be fine with even 1GB because complex detail could be built-up from a minimum of source poly's. Either way, I'm looking forward to seeing instancing, and in Octane 2, hopefully Volumetrics as I hear they're on the way.

Cheers,
Len

Hi guys,

Unfortunately accessing any memory in this way over the PCI-E bus would not just slow octane down 'a bit' and still be useable,
it will simply make octane grind to a halt, eg it will be probably 1000-10.000x slower.

There is no other way than to buy GPUs with more video memory onboard.

Radiance

That's ok radiance, understood, as long as you don't forget the Volumetrics