[JTB]

Just some weeks ago, nobody cared about GPU cards and GPU renderers.. This has changed very quickly and lots of people now know about Octane, Arion, Vray RT, iRay, etc and almost everyone sais that a GPU powered version will be released... (Thea Render roadmap, Indigo roadmap)

Everyone except NL....
So, instead of a 10-15% speed increment that would make us happy for MW 2.5 we ask for a completely new software....
However, this cannot be done easily and soon... (This word has a new meaning in this forum)...

It is vital for many people to know if they have to invest to this new technology, now or next year, with Maxwell or with an other renderer, people that probably want to build 1-2 new PCs for rendering.

I must say that GPU rendering has a long way to go but since new Hardware comes out every day it will be much faster than the development of today's unbiased renderers.

I am confident that NL will be on top again but I am also afraid that they will be a little late... Right now, great small renderers like Octane with an unbelievable price or Indigo which has already a great quality and a devoted users group, will grab a serious number of potential customers...

Unfortunately NL has tried to be like other companies, with newsletters, e-magazines etc... but it seems they had nothing to write about... Now, a newsletter or a public beta or something like other do, is most important...

What do you think? Which will be the next app we get from NL?

[brodie_geers]

I think I'd be happy to know that NL hasn't dismissed GPU rendering out of hand and have/are looking into it.

But I'm not expecting anything anytime soon in that regard. No doubt faster is always better but NL has always focused more on quality than speed and I think there will continue to be a market for that.

Based on what I've seen so far, GPU rendering may well be where everything ends up, but right now the rendering quality just isn't there yet. I'm sure that when NL is able to find a way to produce renders of their current quality, faster than the current engine is able to produce them (whether that's GPU or something else altogether), they'll implement it. They'd be crazy not to from a business stand point.

I will agree with you that current conditions are making hardware purchasing quite risky. The best CPU's on the market might become irrelevant in the near future if GPU has a breakthrough. Likewise, the best video cards on the market may also become irrelevant if that GPU breakthrough involves a technology that's not compatible with current cards.

-Brodie

[Tok_Tok]

Thank you for starting this discussion!

I'm keeping it short; 1. I think, at this point, Next limit has an obligation to inform it's costumers about the near future. 2. If we don't hear anything from Next limit, in this post or by mail, i think we
can assume that GPU won't be availalbe this year.

It maybe good for us to know what we're up to but in the end the last thing Next limit wants to do is scare away their costumers.

On the other hand, Next limit has always been a company with innovative idea and the first with amazing software/features. (Realflow and Maxwell's Multilight) So i'm confident this hasn't slipped their
mind at all.

[JTB]

Well, I can be almost sure that we won't see GPU this year, however, getting a Maxwell2.2 or 2.5 would be a terrible waste of time and effort, don't you think? My guess is that Maxwell will be closer to Arion than Octane (hybrid) and that the price will be also like Arion, making it much more difficult for many of us, considering the additional cost for GPU cards...

I never questioned NL for the quality.... that would be crazy... I just don't want to stay behind especially when speed is a gift from God with the tight deadlines I have to deal...

[Half Life]

GPU rendering seems like a temporary fad to me -- too dependent on specific hardware and I don't see the CPU market sitting still. Right now for about the same price as a decent workstation with a good Quadro card that would be fast enough to make GPU feasible I can get a dual i7(or Xeon) board that will get me 24 cores on a machine that will take as much ram as I can stuff in it and run every other piece of software I have as well. Ultimately that will far out-perform any GPU solution in my daily work.

This seems like a bit of hysteria to me -- "gotta have the newest thing"... relax, wait a little while and another "newest thing" will come along shortly.

Best,
Jason.

[Tok_Tok]

getting a Maxwell2.2 or 2.5 would be a terrible waste of time and effort, don't you think?

At this moment i can't think of any new feature that is more important then GPU rendering. Maxwell users have asked for more speed since forever. So if it doesn't come now, it sure is too late.
I am trying some new engines just to see how it works and which one works best for me. If Maxwell eventually decides to not implement it or a half year from now, i will concidering switching engines
seriously. So, yes, i think they should be focusing on GPU rendering.

Ultimately that will far out-perform any GPU solution in my daily work.

I think that if you buy any CPU and for the same money you get a cuda videocard, the video card always wins, on a time table.

[Half Life]

Only on very specific tasks and I use many different softwares in the course of my day -- weak amounts of video ram are also serious limitations with GPU.

We'll see which of us is right in 3 years though

Best,
Jason.

[zdeno]

RAM issue is rather anything that left to complain about GPU rendering. few years ago when we were on 32bit OS 2 GB RAM was anything we have got and awesome renderings shows. If there will be need then there would show GPU cards with more memory.

in rendering engines market there are only 2 really important factors . speed and quality.
maxwellrender already have awesome quality so I am sure NL is working on speed issue right now (or after football championship in RPA it is their first thing to do on list) (and Yes Bubballo It is football not soccer. That what You have in States should be named American Rugby it is just obvious)

but sadlly we won't know what is cooking until its done. Good thing is. When there was silence from NL updates were splendid, but when there was many roadsmap, newsletters and Nicole's monthly annoucements it always turn to %*$( so just be patient, sigraph is just around the corner

[Fernando Tella]

GPU rendering seems like a temporary fad to me -- too dependent on specific hardware and I don't see the CPU market sitting still. Right now for about the same price as a decent workstation with a good Quadro card that would be fast enough to make GPU feasible I can get a dual i7(or Xeon) board that will get me 24 cores on a machine that will take as much ram as I can stuff in it and run every other piece of software I have as well. Ultimately that will far out-perform any GPU solution in my daily work.

This seems like a bit of hysteria to me -- "gotta have the newest thing"... relax, wait a little while and another "newest thing" will come along shortly.

Best,
Jason.

I super agree. Just wait for some 32 cores chips or so; those are not very far.

What I do like is the instant feedback that usually comes with the GPU solutions.

[Fernando Tella]

(and Yes Bubballo It is football not soccer. That what You have in States should be named American Rugby it is just obvious)

I agree with that too.

[Tok_Tok]

We'll see which of us is right in 3 years though

haha agreed! we'll see where we strand.

(and Yes Bubballo It is football not soccer. That what You have in States should be named American Rugby it is just obvious)

Very true!

[Bubbaloo]

Let's see...

GPU rendering seems like a temporary fad to me -- too dependent on specific hardware and I don't see the CPU market sitting still. Right now for about the same price as a decent workstation with a good Quadro card that would be fast enough to make GPU feasible I can get a dual i7(or Xeon) board that will get me 24 cores on a machine that will take as much ram as I can stuff in it and run every other piece of software I have as well. Ultimately that will far out-perform any GPU solution in my daily work.

This seems like a bit of hysteria to me -- "gotta have the newest thing"... relax, wait a little while and another "newest thing" will come along shortly.

Agree.

When there was silence from NL updates were splendid

Agree.

Just wait for some 32 cores chips or so; those are not very far. What I do like is the instant feedback that usually comes with the GPU solutions.

Agree.

1. I think, at this point, Next limit has an obligation to inform it's costumers about the near future. 2. If we don't hear anything from Next limit, in this post or by mail, i think we
can assume that GPU won't be availalbe this year.

I disagree.

No doubt faster is always better but NL has always focused more on quality than speed and I think there will continue to be a market for that.

Agree.

At this moment i can't think of any new feature that is more important then GPU rendering. Maxwell users have asked for more speed since forever. So if it doesn't come now, it sure is too late.

Disagree.

and Yes Bubballo It is football not soccer.

Whatever you call it, it's a bore.

[Maximus3D]

I have a feeling we will not see a GPU supported Maxwell engine anytime soon, as 2.0 is relatively fresh they are probably focusing their efforts on that to improve it, make it better and integrate the plugins even better with new plugins added over time for new applications as we have seen started already with Photoshop and others.
They are probably keeping an eye on GPU rendering to see in what direction it goes and what happens but my guess is they don't wanna jump on the wagon too early incase things change radically for the better soon and the current exisitng technology will be outdated.

That's my guess.

/ Max

[zdeno]

Whatever you call it, it's a bore.

agree

I prefer basketball . Go CELTICS !! but in Madrid it is something like religion so I think NL have it's holiday right now. Maybe Nicole and others are in RPA right now ?

and about bore - didn't You see baseball ? That is boring x5. But plenty of time to eat burger or watch some commericials.

BTW american rugby is a great sport to watch during CG job. 10 minutes of setting plays and then 10 seconds of action.

and about GPU and others. the best would be if there are better CPUs with all this multicores as someone said here.

[Bubbaloo]

I agree baseball is very boring too. America's pastime, nope. The only sport I make it a point to watch is college football, specifically the University of Oklahoma. I like to watch tennis too. But any sport where an "exciting" game ends in a score of 1 - 0?

Wait. What were we talking about?