Don't Die XSI

[Pancho]

If you go for 5 cards: Which case, power supply, board? Or are you going to put those 5 in 5 PCs?

[jamination]

5 is probably pushing it, the returns diminish with each card, I believe the sweet spot is 2, a lot of people use four. I do wonder how they get them in there though, as I think most of these cards take up 2 slots.

[Helli]

Did you try that with more then 2 cards ?

Tried Octane with 3 Titans and it actually scales linear, there is no loss at all with more cards.
In the end the time the scene needs to be loaded into the card might allways be the bottleneck, but the actual rendering should scale quite well.

[markraz]

GPU aside, and as far as general quality goes, is this redshift better than mentalray, vray and maxwell etc... ??

also have you guys used Arnold?

[julius]

Redshift is a very good renderer, and can render with same quality than any other "big standard" renderer.
It's easy to use when you come from MR or Vray and a lot faster. The performance scale isn't linear with multiple cards when all of them are rendering the same frame. But you can attribute a different frame for each, and then you'll multiply your perfs by your number of cards.

[FXDude]

Below is a little addendum to something posted on the list : RE: maya graph dependencies

______________________________
Shake was EOL'ed in mid 2006

2007 ...
...
2008 ...

...

and finally, ...
2009...

Is Nuke the new Shake?
http://provideocoalition.com/mchristiansen/story/is_nuke_the_new_shake

Most studios did not feel compelled to make a sudden investment in new software at that time as there was substantial investment in a Shake pipeline at many facilities.

Among the alternatives from Autodesk, Avid, Adobe and Eyeon, there was no clear successor.

So of what could we say "Is X the new XSI?"

I see people trying to do Ice stuff in Modo, wrap their heads around, or produce stuff fast in Maya or Houdini... that take "2 seconds" or are no brainers in XSI,
use ten different things importing back and forth to do 10 different things...

Perhaps -someday- Houdini, Blender, or if Maya got humanized to the point of not being recognized...
or it seems that Premo has been described not unlike "a new XSI" already today if it would ever get released.



But whatever, or for whenever this X would be,... is that day today?... but I mean like at-all?

(Notwithstanding the incredible flexibility of Houdini for example :: Can you work quickly in Houdini? )


So why would we be bending ourselves in 4 and move before then. For Autodesk?

Perhaps it's because of how things were handled (probably not by mere circumstance), but I guess I'm also upset at how easy it was to get to this point.

New Softimage development is sometimes barely given the time of day on the (sometimes everything but) Softimage forum, and where is everybody? ... Blogs get abandonned..
that while seeing people making backflips to compensate with "Toxik", "AfterEffects", or "Final Cut Pro".

At least some studios that have to deal with daily deadlines (and perhaps Chinese competition) decided to go-on working with what works best for them
(while also producing awesome stuff you'd have a hard time thinking with what else they could have done that, in that given time, or with that amount of artists .. and perhaps a few 'technicians'... working)

But I must admit that I would have hoped that such reasoning would have been somewhat more widely held, but that at least makes for some fair Softimage (post-)life.



A shake Forum slightly before activity/usage went down (some ~4 years post EOL),
just about when (and -only- when) something became a (much) closer alternative to the quite exceptional nature of Shake. (also very advanced while equally user friendly)

[MauricioPC]

The only difference I see is that Shake was widely used, so you could find a lot of work with Shake. Softimage houses aren't great in numbers, which is a shame.

I would gladly study Softimage again instead of Maya.

[FXDude]

yes fair points...

Perhaps also would there be some feedback loop effect going-on there.