Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - lolec

Pages: [1] 2 3 ... 12
1
Off-Topic / Re: Here we go
« on: 2022-09-26, 20:06:42 »
Oh yeah my next question was going to be about neural rendering :P are you guys playing around with that?

2
Off-Topic / Re: Here we go
« on: 2022-09-26, 19:56:51 »
BTW. I realize this question often comes from the OMG FSTORM WOW OCTANE OMFG crowd.

That is not where I'm coming from. If we compare the best CPU available in 2013 to the best available today we get around a 12x increase in render speed. If we do the same for GPU its about 6x, that's why I think it was such a good decision, but we still don't have info for the 4090 so the gap might close this year.

3
Off-Topic / Re: Here we go
« on: 2022-09-26, 19:34:40 »
Yes, I understand the limitations and I also believe it was the right decision when it was made in ~2013?  But I think it is wise to challenge assumptions from time to time. I absolutely don't have all of the information, you guys are the genius minds who made the right decision in 2013 when you had all of the information, but just because things haven't changed in the past that doesn't mean they won't ever change.

Is the market similar enough that the decision still holds? I truly don't know! That's why I'm asking and I trust your answer. I can only see one side of the picture and that is market pressure. There is enormous market pressure for better and better GPUs, Gaming, AI, even crypto etc.

But the market pressure for better CPUs is not that strong, the evidence is how we have been essentially stuck at the high end for a few years now.

I don't think this will flip the market in the near term, Corona is and will continue to be one of the most popular engines! But stay humble and nimble! re-evaluate your position constantly! don't let a shift in the industry take you by surprise as it happens with so many companies! For every example of an industry not changing I can give you one where the industry changed suddenly and the leaders didn't react on time. (kodak, blackberry...) I love you and don't want that to happen :)


4
Off-Topic / Here we go
« on: 2022-09-26, 16:57:20 »
Ok, I know this is a heated subject and that it has come up several times in the past, however, I think it's important to re-evaluate our assumptions from time to time, specially when initial conditions change so drastically.

While I understand and agree that 8 years ago the right decision was to be proudly CPU only, it seems like we live in a completely different world now.

1. High end GPUs have large amounts of VRAM
2. GPU compute power is growing at an insane rate compared do CPU
3. Rendering tech also evolved to make GPU more capable and versatile
4. Vray is Hybrid, Corona is part of Chaos now.

So my question is, Has the situation changed enough to make you consider changing your original stance and maybe work on a Hybrid solution to take advantage of the good that GPU has to offer? Is that even technically feasible or is Corona so strictly CPU only that there is no way to take advantage of that other than denoising?

I fear that at the current rate of progress in hadrware, Corona won't be able to compete. I love Corona and want it to continue to be the best render engine forever :)

 

5
[Max] I need help! / Distribuited Render Issues
« on: 2022-07-13, 22:22:51 »
Hello, I'm new to distributed rendering. I Just built a small cluster of 7 5950x machines, 6 of them have 32gb of ram and one has 64 gb. My workstation is 3990x with 256gb of ram.

I started running tests and I found something unusual, maybe someone with more experience can tell me why something like this happens.

I sent the render to only 2 nodes, one with 64 and the other with 32 of ram. AS the render progresses, the one with 64g is working properly, but the other one is not contributing any passes, even though the processor it's been at 100% usage for 30 minutes or so.

Both have the same processor and joined from the beginning,  any ideas ?

Thanks.



6
Hardware / Best bang for buck render node.
« on: 2022-07-05, 07:15:43 »
Hello!

We just got a big project that will require a lot of animation in the next 12 months or so. We want to build a small render farm.

My workstation is a threadripper 3990x 256 gb ram ( although I rarely use more than 32 GB for rendering) . I would like to get around 5X my current rendering capability. But now, I'm faced with a question I've never faced before. As building a rendering workstation is a relatively easy decision, but I'm finding building nodes is not as straightforward.

This is because the case,  power supply, cooling, GPU, motherboard, storage and even ram scale with the number of nodes, so apparently the equation would lean towards minimizing the number of nodes, but at some point, the performance per dollar plummets so buying 4 3990x based systems might not be ideal.

I'm finding it hard to figure out what the absolute best strategy is.

Is there a currently accepted "best" node configuration?

Thanks :)

7
I rarely need to do animation so I'm a noob at it. I've been struggling to solve this for almost a week, hopefully someone can give me a tip that unlocks my brain.

Here is what I need to animate in max:
https://share.getcloudapp.com/nOu9ZPn0


Here is what I got so far doing manual animation (the green cylinders are supposed to overlap at all times.):
https://share.getcloudapp.com/yAukErYQ


Manually animating both groups to match is impossible, so i'm thinking on creating some kind of rig that locks the green cylinders together. but I have no idea on how to do that.

Anything will be appreciated :)


8
Hello people,
I was wondering is there AMD CPU available that can turn Corona Interactive into "real-time renderer"  (whenever I change something in the scene, Interactive Render turns into blocky mess for second or two before image is cleared) . How close are Threadripper 3990X and similar powerhouses to this?

It depends on your scene. But for simple scenes. Pretty close.

9
General CG Discussion / Re: Object library ?
« on: 2021-08-21, 19:41:14 »
Thanks for the links, I didn't knew about some of these :) 

10
wait, what about decals? Was this added already, how did I miss that?
There might or might not be something done, but sadly because some of the users cannot understand that a feature might get postponed when unexpected complications occur, we will generally announce only features that are fully done ;)

Yeah, I feared that would happen damn.

If you ever open an "insider" program with a 0 tolerance for bitching and tantrums, count me in.

11
General CG Discussion / Object library ?
« on: 2021-08-03, 04:39:19 »
I'm looking for a large high-quality 3d asset library.

We already use Megascans, Poliigon and Dimensiva (plus random assets from turbosquid etc).

I'm more interested in the type of 3d assets that Megascans has. Is there another large library of 3d assets like that?

Thanks.

12
[Max] General Discussion / How Can I Help ???
« on: 2021-06-11, 22:07:27 »
Corona Renderer is a cornerstone of my work. Although I don't work in a render-heavy industry like ArchViZ or Animation, Corona allowed me and my team to do things we couldn't do before. I cannot share the exact work we do, but suffice is to say that Corona has provided the greatest ROI of any tool I can think of. By several orders of magnitude.

When I started my company 5 years ago, I was able to start selling my work before it existed because the images looked like actual photos. I didn't have the money to build the actual products, so this changed the course of my life dramatically.

It's not like I started from scratch, I had been using Maxwell renderer and Vray for years. But Corona resonated with my brain and I suddenly was able to do much more.

Anyway, my work involves a lot of product development and I go through releases 4 - 8 times a year. I've seen the Corona team go through a lot of what I go through and sometimes I feel their frustration as if it was my own. The recent discussion about V7 and V8 has made me realize how much Corona has done for me and I just want to give back a little bit.

Other than the standard ways in which us users can help (testing builds, helping new users etc...) I would like to know if there is anything else I (we) can do to help! I bet there is a lot of other people here that would like to help in some ways too.  I post this publicly in case other people want to offer their help too, but the best ways to help are probably private. And in any case, thanks for building Corona!


13
That comment above about the house and the city is so flawed.
If you gonna start building a city you better hire a construction company, not rely on the same three people that did your house...
Because news flash those three people will do a terrible job and it will involve a lot of problems and a lot of setbacks, also they will promise you deadlines they can not keep. Sound familiar?

I struggle to understand how someone could be so out of touch with reality that they will tell the people actually DOING the thing "what you say is flawed"  How would you know more than them? Can you imagine I come to your work and after looking around for 10 minutes I start telling you that what you are doing is wrong? After you've spent years doing that?

The problem with analogies is that they only work on people who already agree with the premise in the first place but criticizing the analogy is absurd... Yes software is also not made of bricks... yes, software doesn't need streetlights.  Obviously, corona 1 is not EXACTLY like a house and corona 7 is not EXACTLY like a city.  But as Tom said, Corona 1 had to worry about much less things than Corona 7.

I hate forum arguments, I know they don't convince anyone, I know I look as much as a fool as the people who disagree with me. I know I look like a mindless fanboy.

I'm not doing this to convince anyone.

I know exactly what it takes to build products, I know how hard it is and discouraging it is to read comments like the ones we've seen in the forum lately. I'm also very familiar with the frustration of working your ass off for a year and not being able to say exactly what you think because as creators, you need to keep composture.

Reading some of Ondra, Maru and Tom comments, it's clear that the negativity is getting to them, enough for them to break character and express their disappointment.

My only goal is to offset the negativity and make the dev team feel a little less crappy.

Corona is a fucking brilliant piece of software. I see and appreciate the little details that go into that.  A product is much more about what It DOESN'T have than what it DOES HAVE. Corona is brilliant at not adding stuff, I know this will only be understood by a handful of people and will be seen as crazy talk by the rest. Hopefully, the dev team will smile knowing that every time I use Corona, I appreciate all the things it doesn't have. I even appreciate that it doesn't have Tonemapping YET, because doing great things take time. 

Many people say they decided to move to Corona because of the features mentioned in the roadmap, not only do I highly doubt that, but also it's a terrible decision. Personally, I didn't choose Corona because of the current or future features, but the ideology and principles behind it.  These principles guide all future decisions. The same people that decided to do all the things you love about Corona, decided to move tonemapping to V8, it is reasonable to expect that you will end up loving that decision too. Just be patient.

Thanks again for the decisions you make, I hope you have the resources you need to fulfill your vision and that you don't get discouraged. Thanks for making hard decisions, this shows real love for the product, and when a product is made with love, it makes a difference.

Looking forward to v8!






14
I think there is a discussion to be had about whether the new Physical Material was worth the dev team required to implement. I know that sounds harsh, but I have not yet found that it significantly improves my shading workflow - I feel like I am deciphering it more than I am wielding it.

Unless you are willing to spend the time they spent discussing the pros and cons, willing to study and understand the industry as well as they do, and most importantly... unless you are willing to suffer the consequences of the decision... NO, you have absolutely no place in that discussion.

If this was a bad decision, they will suffer the consequence. People will stop using their product. You are not even willing to spend the time to understand how to use it, while they spend MONTHS working on it because they think its important, based on information you and I don't have. 

15
The message explaining the situation is perfect. Completely understandable. It looks like the team is doing the important and not glamorous work that makes Corona great. Sure it takes 15 minuts for the Fstorm guy to implement new flashy features.... how long has it taken him to make a robust production-ready renderer? right... hasn't happened yet.

It takes courage and conviction from the dev team to work on something "no one was asking for" knowing it will make the product much better in the long term.

Again, I think the only mistake was not explaining this before editing the Trello. After reading about how quality and doing the hard work, even knowing people will be mad temporarily, is the most important thing, I feel even more confident that the new tone mapping will be great and I want to spend a little time sending my support and encouragement to compensate some of the comments from people who don't understand the process.

Anyone comparing Corona to buying a product and not getting what you paid for are just trolls. You should pay for software based on the current stable version. If you are paying for a software based on features that are written on a Trello called TENTATIVE ROADMAP, it's completely your fault.   




Pages: [1] 2 3 ... 12