Login - Create Account
Bookmark and Share

Is FF Inc. making something for FF 4.0 to have a FASTER render engine?

<< 1 2 >>

Login or Register to post new topics or replies
uberzev
not uberfez

Posts: 1827
Filters: 36
Vlad is the big cheese at FF and I'm sure he could address some of these concerns.

I can bug him if you guys would like me to...
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
As Skybase wrote brilliantly and very well and I agree totally

Quote
I should quit posting on this stuff.

Rather see art now

Yup.

Going back to doing art.

Art.

Doing Art more important than continue arguing here


AND more specially after someone say very stupid and silly things as:
"About 99% of the filters in the library are completely useless.." smile:evil:

How can someone be able to say this ???? smile:?: smile:?: smile:?:

Will not continue on this unless the FF team gives some news about this and I doubt that the FF team will give any news about this OR worse will make any change in the rendering engine to make it faster, so I think that all this is useless. Sorry

Today GMM will see this thread and will not care at all to give any information here and will be like many other days that this thread is invisible to the FF team.
  Details E-Mail
uberzev
not uberfez

Posts: 1827
Filters: 36
Quote
"About 99% of the filters in the library are completely useless.." smile:evil:
Agreed
  Details E-Mail
Morgantao
Can't script

Posts: 2185
Filters: 20
Quote
Today GMM will see this thread and will not care at all to give any information

I think you'r barking at the wrong tree, SpaceRay.
GMM is here to help us with anythingn he can, but he's not a policy director. If the FF big anchiladas tell him not to release any info about coming features, he won't release any information. smile;)
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
Sorry, I am not going in any way against GMM, he is good and help us as much as he can, BUT the problem is that he must follow the FF team rules, and so he CAN´T do what he wants, so if he can´t help us on this topic of this thread because the FF team does not give him the needed information or what to do, he is forced to act like this thread is invisible and do not exist, isnot his fault, is from the FF team.
  Details E-Mail
Skybase
2D/3D Generalist

Posts: 2822
Filters: 67
lol what are you trying to do? Make people leak info?
  Details E-Mail
uberzev
not uberfez

Posts: 1827
Filters: 36
SpaceRay is looking for his Deep Throat.
  Details E-Mail
Sign Guy
Digital Art Developer-Publisher

Posts: 358
Quote

SpaceRay Wrote:
Sorry, I am not going in any way against GMM, he is good and help us as much as he can, BUT the problem is that he must follow the FF team rules, and so he CAN´T do what he wants, so if he can´t help us on this topic of this thread because the FF team does not give him the needed information or what to do, he is forced to act like this thread is invisible and do not exist, isnot his fault, is from the FF team.


I don't see anyone speaking against any individual such as GMM going on. I do see discussion of the shortcomings of the Filter Forge program and the management decisions of the FF team with regards to these shortcomings. Whether they choose to respond with replies here or with actions on upcoming releases is up to them.

It's all about word of mouth advertising. When people here and elsewhere post good things about FF, that spreads quickly and benefits sales. When people here and elsewhere post complaints about FF, that spreads quickly and hurts sales. Thus, it becomes a business decision as to how, or if, to respond for the management of Filter Forge. It certainly isn't personal.

As a paying customer for the Filter Forge software, each of us is clearly acting in our own best interests and the interests of all customers by giving voice to such matters.
Fred Weiss
Allied Computer Graphics, Inc.
  Details E-Mail
StevieJ
Designer/Artist

Posts: 11251
Filters: 163
Interesting discussion...

Since FF is procedural, a software accelerator is possible...but unlikely FF will invest in the programming needed to create one... Think FF will remain the same...and more likely to make adjustments to take advantage of new and faster processors...and more cores in the future...

+1 ...on no quality control...and wayyyyy too many useless filters... FF wants the number of filters as a selling point...obviously...

+1 ...on using FF to create base images with Substance Designer...excellent complimentary combo...

+1 ...on positive and negative feedback affecting sales...guilty of alot of negatives myself...and has been for things towards my own self-interests...but also think my "complaints" would benefit FF if as well...if they were ever addressed... smile:|

+1 ...on sending Uber to bug them... smile:dgrin:
Steve

"Buzzards gotta eat...same as worms..." - Clint :)
  Details E-Mail
Kraellin
Kraellin

Posts: 12603
Filters: 97
i remember a similar discussion on another subject within FF a while back. i recall that vladimir talked about the costs of program libraries (.dll's) and that those were one of the biggest reasons they didnt or couldnt do whatever that other discussion was about. i'm fairly sure the same would apply here. some of the .dll's one might use in one's program can be QUITE expensive. building them yourselves can also be time consuming and expensive. thus, one tends to use more cost effective libraries. i'm going to guess that that same arguement would apply to this present conversation. we may not like it, but in order to keep selling prices down and development ongoing at a reasonable pace, the 'best' and fastest programming is not always within one's budget.

now, i dont know that that's the case here. i'm just bringing up the point that it may be, based on that prior conversation on whatever that other program idea was smile:)
If wishes were horses... there'd be a whole lot of horse crap to clean up!

Craig
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
Quote
Skybase wrote:

lol what are you trying to do? Make people leak info?


NO, just a very simple and very short word

YES or NO

Do not want to get the details, just the answer to question of this thread.

If you read well the question on the title I am not asking how they are going to make it or get more details about it

JUST SIMPLY YES OR NO

If the answer is NO, anything else is totally useless.
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
As you have seen GMM has answered to others threads but as I said ignored this one forced surely by the FF team that does not allow him to give any information on this thread.

As I have said above, I do not want MUCH information, ONLY AND JUST a simply YES OR NO
  Details E-Mail
Skybase
2D/3D Generalist

Posts: 2822
Filters: 67
lol ok spaceray. ok. smile:p
  Details E-Mail
uberzev
not uberfez

Posts: 1827
Filters: 36
Would be a shame for GMM to reply now. smile:D
  Details E-Mail
Morgantao
Can't script

Posts: 2185
Filters: 20
smile:D smile:D
  Details E-Mail
StevieJ
Designer/Artist

Posts: 11251
Filters: 163
Feel a "SpaceRay vs GMM" cage match is in order here... smile:D lol...
Steve

"Buzzards gotta eat...same as worms..." - Clint :)
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
smile:D There is no fight because there is no oponent, as GMM will not answer in this thread, and even if it happens that miracle, I would not fight at all, just ask for a YES or NO.

I woud bet and think that the most possible answer is NO, Why? Because if the FF team avoids to answer in this thread is because they are not doing something to make FF faster, or they want to do it but can´t do it because in the way FF was built, and so they can´t give an answer, and if they say yes and can´t really make it would be bad.

If it was clear that they ARE doing something positive and reaching good results for making a faster FF 4.0 they would not have a problem to say a YES, but they are hiding from any answer, so most probably is NO BUT they do not want to say it so there is still the doubt and possible hope.
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
9 days more without any news from the FF team, and I only want to know if they have any possible plan to make FF 4.0 faster when rising the resolution of the results.

JUST AND ONLY want to know a YES or NO, do not need to explain or write more. Or at least we are trying to do it.
  Details E-Mail
Skybase
2D/3D Generalist

Posts: 2822
Filters: 67
Darn, I waited 9 days for a response! Man I get so much nothing on this forum! LOL. smile;) One of these days. ahaha.
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
smile:D smile;)

Yes, so much nothing from the FF developers team smile;) that does not give any possible answer.
  Details E-Mail
joBox
joBox

Posts: 21
Filters: 3
There is ONE problem with Substance Designer that is an absolute deal breaker: The largest image that can be rendered is 4096x4096.

http://support.allegorithmic.com/docs...Dialog.htm

Scroll to bottom to see:
"Cooking size limit" defines the maximum pixel resolution permitted. By default, this is 4096 pixels. As a Substance's output is always square, the value set here defines both the maximum width and height, in pixels.
  Details E-Mail
joBox
joBox

Posts: 21
Filters: 3
Oh and Genetica is $899 and can only render up to 6000x6000
http://www.spiralgraphics.biz/gen2tou....htm#tour1
  Details E-Mail
Morgantao
Can't script

Posts: 2185
Filters: 20
I guess the saying "Nobody's perfect" goes for software too smile:)
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
Quote
Jobox wrote:

Oh and Genetica is $899 and can only render up to 6000x6000


I have to say that this price of 899$ is WRONG if you want to compare Genetica with Filter Forge, as is not true at all, because you should choose the Pro version and NOT the Studio version that have animation.

Genetica Pro is the nearest posible comparison to Filter Forge and is 399$

Although really Filter Forge is cheaper as it has many discounts usually and Genetica has ONLY as much as 30% once in a while very few times.

See here below the difference

CAN´T COMPARE REALLY GENETICA TO FILTER FORGE

Although both are node based and can possibly make similar things, they are really in my opinion and point of view very different and can´t really be compared against each other.

  Details E-Mail
Vladimir Golovin
Administrator
Filter Forge, Inc.
Posts: 3147
Filters: 55
Quote
Please, I want to know if the FF Inc. is working on the development of a optimized and better render engine that could process the filters faster than what is available in FF 3.0.


Alas, the situation on this front hasn't changed much. The most viable way to speed up the renderer at this point would be to rewrite it to use a GPU-supporting architecture, but the existing codebase is built around the CPU paradigm, so a straight port is extremely difficult. I posted about this before.

We experimented with GPUs, observed insane gains in performance on some tasks, but we decided not to pursue this approach. Basically it would mean rewriting the renderer from scratch, plus we'd likely encounter big, really big problems with backward compatibility.

If we started writing Filter Forge today, or a couple of years ago, we'd certainly build it to utilize GPUs. This would also work for a separate product that doesn't have to maintain compatibility with Filter Forge (that is, being able to render its filters from all versions down to v 1.0).
  Details E-Mail
Vladimir Golovin
Administrator
Filter Forge, Inc.
Posts: 3147
Filters: 55
On the other hand, there have been some interesting developments on the GPU front that are relevant to Filter Forge. For example, id Software's RAGE suggests that we can use an approach similar to their megatexture streaming to implement Filter Forge's infinite bitmap caches. And Epic's UE4 utilizes GPU-accessible data structures such as octrees to do their fantastic global illumination.

Which means, we probably should revisit the GPU idea.
  Details E-Mail
Skybase
2D/3D Generalist

Posts: 2822
Filters: 67
High five. smile:D
  Details E-Mail
Kraellin
Kraellin

Posts: 12603
Filters: 97
thanks, vladimir smile:)
If wishes were horses... there'd be a whole lot of horse crap to clean up!

Craig
  Details E-Mail
Morgantao
Can't script

Posts: 2185
Filters: 20
Good to hear that there may be some use of GPUs in the future. smile8)

I would like to thank you for taking the time to explain some of the complexities of adding GPU support. It's important to hear fr om the team wh ere things are going smile:)
  Details E-Mail
Vladimir Golovin
Administrator
Filter Forge, Inc.
Posts: 3147
Filters: 55
There's another potential way to speed up the rendering: the new Xeon Phi co-processor from Intel (formerly known as Larrabee):
http://www.anandtech.com/show/6017/in...es-retail/

Implementing support for it shouldn't be too hard. However, unless this card becomes a must-have upgrade for graphic artists, I don't think it's worth the effort.
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
Thanks very much for answering in this thread and good to know that there is possibly some slightly hope on GPU acceleration.

Although I thank you for answering and giving all this comments you have put, I am sorry to bother you and say that you really have NOT answered to the question in the title of this thread.

Please, could you tell if there is any possible plan about this for FILTER FORGE 4.0 ?

Is FF Inc. making something for FF 4.0 to have a FASTER render engine?

I am NOT speaking here about GPU at all, I mean ANY way that you could think of to make FF 4.0 faster than FF 3.0
  Details E-Mail
Sign Guy
Digital Art Developer-Publisher

Posts: 358
For example, the suggestion of a global on/off setting in preferences for anti-aliasing.
Fred Weiss
Allied Computer Graphics, Inc.
  Details E-Mail
Skybase
2D/3D Generalist

Posts: 2822
Filters: 67
lolol where's the popcorn?!! I'll be sitting in the back row enjoying the thread. lolol

Shhhhh patients spaceray, 4.0 is around the block. We'll find out soon enough.
  Details E-Mail
Morgantao
Can't script

Posts: 2185
Filters: 20
I forbid FF 4.0 beta coming out, untill I have enough HU points to upgrade!
So go ahead and start using my filters like crazy smile:-p
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
Quote
Morgantao wrote:

I forbid FF 4.0 beta coming out, untill I have enough HU points to upgrade!


Since the release of the Beta there has been between 6 and 9 months until the final release version, so probably if your filters are used much you could have time to get the HU points, OR even better make a new (or more)awesome and very useful filter that many people would love to use and get those wanted HU rewards

see this thread for more information of the release dates of the betas and final versions

http://www.filterforge.com/forum/read...5&TID=9638
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
Quote
Skybase wrote:

lolol where's the popcorn?!! I'll be sitting in the back row enjoying the thread. lolol

Shhhhh patients spaceray, 4.0 is around the block. We'll find out soon enough.


smile:D smile:D smile:D Glad you enjoy it smile:D smile:D smile:D

Here you have some FF Popcorn smile:D

There is something very important, is supossed that FF 4.0 will be released early in next year 2013, so if this possible way to make a GPU acceleration that Vladimir have already told above is NOT included and will be something for FF 5.0, Filter Forge Inc. will be loosing much against the competing software companies that are ALREADY NOW making updated faster versions of many Photoshop plugins and also some others software companies are going to release even better versions at the end of this year.

So if FF Inc. is NOT going to make FF 4.0 faster and have the SAME render engine of FF 3.0 will be a great loose. WHY ? Because FF 5.0 will be not available until probably the middle or end of 2014 and already at the middle of 2013 there will be MUCH better plugins and much faster ones available and FF will be left behind

I know that there is really no other direct competition software available, BUT there are many available plugins and artistic softwares that CAN make many of the effects and things that FF can do, and MUCH FASTER AT HIGH RESOLUTIONS AND WITH HIGH QUALITY than FF 3.0 can do.

I LOVE MUCH FILTER FORGE and would not like to see it left behind against other companies, and I will not continue using Filter Forge in the same way and will regretably have to use OTHER software that are really optimized and faster instead of FF.

AND VERY IMPORTANT THAT MANY OF THE PLUGIN SOFTWARES I KNOW DOES NOT USE GPU ACCELERATION TO MAKE IT FASTER, JUST OPTIMIZED RENDER ENGINE OR CODE
  Details E-Mail
Skybase
2D/3D Generalist

Posts: 2822
Filters: 67
I mean, this is actually a really really deep topic of how you output data in a visual form. The current blog article I'm writing (which is longer than I imagined) touches on topics regarding rendering in the most general manner possible. Hopefully some of the little things I'll go over will be helpful in saving tons of time from a single render.

And SpaceRay, I highly doubt there will be something that can beat FilterForge in the next couple years. This is one solid piece of software that has more functionality than anything I ever used in a long long time. Back in the old days KAI's Power Tools was "the thing" for the Photoshop market now the market's mildly shifted but this software's basically "the thing." There's something special about this software and maybe you haven't seen the crazier things you can achieve with this software that most others won't even deal with.

See the thing is, you'll see software that's FASTER than this program, but when it comes to the amount of stuff you can do, FilterForge usually wins by a long shot. smile;) And I mean it, I've dug through hundreds of similar, procedural-networking graphics tools out there and while I continue to use many of them, FilterForge is the number one weapon of choice when it comes to rapid tool development or image generation. Sure there are tons of things this program can't do but it's much much better and easier than the rest of the gang.
  Details E-Mail
Vladimir Golovin
Administrator
Filter Forge, Inc.
Posts: 3147
Filters: 55
Quote
Is FF Inc. making something for FF 4.0 to have a FASTER render engine?


For FF4.0, no. The current CPU-based renderer and most components are already optimized to their limit, so trying to optimize them further will push us beyond the point of diminishing returns. All possible improvements I mentioned in this thread are long-term plans.

Quote
AND VERY IMPORTANT THAT MANY OF THE PLUGIN SOFTWARES I KNOW DOES NOT USE GPU ACCELERATION TO MAKE IT FASTER, JUST OPTIMIZED RENDER ENGINE OR CODE.


Filter Forge filters will never be able to offer better speed than hard-coded, precompiled filters because FF uses a node-based architecture that imposes a runtime cost (extra function calls, passing parameters etc). Also, Filter Forge filters don't exist at compile time, so they cannot benefit from whole-program optimizations offered by modern compilers. It's a deliberate tradeoff: either you get better speed, or you get node-based flexibility.

Also, we use double-precision for almost everything, while other software makers can use integer math and buffers when necessary. Integer calculations are cheaper than double-precision floating point calculations, but we can't get rid of them, again due to our node-based architecture.

So, you shouldn't expect Filter Forge to render some effect faster than a hard-coded, compiled version of the same effect on the same hardware.
  Details E-Mail
SpaceRay
SpaceRay

Posts: 8260
Filters: 33
Quote
Vladimir Golovin wrote:

For FF4.0, no.


WOW!! I finally after so much time I have a real and useful answer so I do not keep bothering and annoying all the forum with the same topic

You see, it was not so difficult to answer to the title question.

Now I know it and will not keep asking for it anymore.

Thanks very much for telling it and also explaining why, very good to know.

Quote
Vladimir Golovin wrote:

The current CPU-based renderer and most components are already optimized to their limit, so trying to optimize them further will push us beyond the point of diminishing returns. All possible improvements I mentioned in this thread are long-term plans.


Thanks and good to know that we can´t expect any kind of speed optimization of the render speed as it is already maximum optimized.
  Details E-Mail

<< 1 2 >>

Join Our Community!

Filter Forge has a thriving, vibrant, knowledgeable user community. Feel free to join us and have fun!

17,426 Registered Users
+6 new last day!

121,237 Posts
+10 new last day!

11,138 Topics
+15 new in 7 days!

Online Users Last 5 minutes:

hayssm, 5 unregistered users.

Recent Wiki Edits:

Follow filterforge on Twitter