New Release Candidate coming this week?...

[johlanga]

2-8x used gtx 580(or 4x gtx590) cost almost nothing...

And the pros could buy 8x Titans

+1
Discontinuing pre-Fermi GPU suport.

[FrankPooleFloating]

+1

Kill 32bit and pre-Fermi cards. As you see, these slowed devs down... I point out
in something like 10 threads, my 460's cost me $190 total for both (672 Fermi Cores)...
if you have a 260 or something, time to hop on eBay or Craigslist.

[necko77]

+1
Discontinuing pre-Fermi GPU suport.

Octane Team please dont waste your time developing software for hardware which is replaced in almost every pc...

Think about tomorrow...

[FrankPooleFloating]

Hey Marcus (or other powers that be)

What are the top new features we should be expecting on Monday or Tuesday? (provided
everything goes okay)... all this stuff is spread across so many threads. I can't even begin
to remember exactly what all is coming in next version... I want to be excited about what
is coming... hard to do when you don't know what that is.

It didn't sound like Undo, Background Image, Primary Visibility or Sun Size was coming,
if I remember correctly...

You did say this about Sun Size:

Re: Daylight System Shadows
We can fairly easily add a new pin to change it in the daylight node. I assume that you want to keep the total brightness of the sun the same.

If this could be squeaked in (because it is an easy one) that would rock indeed.

(crap - I just remembered it is like midnight Friday there... oh well.. answer this weekend
if you get a chance)

[pixelrush]

I think Octane by nature is a 64bit program.
It needs losts of ram for use but also by the sounds of it for coding.
I have a 32bit boot on one PC for older but still perfectly useful apps but they are really also becoming outmoded on 32 bit. I have Octane installed there as well to render simple scenes for product viz during the concepts phase and for that 32 bit is fine.
It would not be too bad if development of 32 bit stopped here or soon however.
The card I have on that PC is a 460 so I wouldn't be personally affected immediately I just thought to speak up here in the testing forum for those that might. If coding for older compute capability has become a problem/limitation its best to move on. It doesn't seem like this redundancy will happen too often as cuda matures. If Otoy could wring 32 bit builds out for post 2xx cards and 64 bit for pre 2xx until the end of v1 that would be nice but its not something they should spend weeks trying to bring about.
I have just built a new 64 bit multi gpu rig and that's where I intend to do most rendering in the future anyway.

[mainframefx]

Regarding 1.11: It looks as if everything is sorted out for the release on Monday/Tuesday. Building for pre-Fermi GPUs is a pain though and it may be that this is the first release without support for them I can't say yet for sure.

I gotta say I dont see a huge issue with that. There were a lot of architecture changes, this was only the second generation of CUDA cards, a quite young technology back then. I think the Fermi+ series aren't running into any trouble within atleast the next 2-3 years. You can get very cheap Fermi+ cards that perform quite a lot better. So I don't see a lot of financial issues there.