Sorry, what I meant was, is there a list of some of the things you did to achieve the “1.7 to 2.4x faster rendering with CUDA”?
@bliblubli Can you show a link to your discussions with Blender Institute to incorporate your changes into Master? I’m sceptical that they would refuse to merge something which performs much better unless it breaks other things in Cycles such as importance sampling or messes with results like scrambling distance does. The devs aren’t unreasonable, and for those of us who understand what implications certain changes might have, it is understandable that they might not want to merge whatever changes you have made.
I’m sure you’ve done some great work on improvements in Cycles, but I feel there are better ways to benefit from contributions to open source software rather than trying to sell your changes. Use your improved efficiency to earn more, use your expertise gained as validation for your programming ability and use it to negotiate a higher wage as a developer, almost anything is better than trying to sell changes to open source software. It only takes one person who feels like releasing it for free and your whole idea is down the drain.
People can benefit from open source software, as many corporations have recently discovered. You just don’t benefit from it by trying to sell it. I commend you for the work you’ve done on Cycles, but I feel like you could have a much better public image if you release what you’ve done and teach others so that everyone can benefit.
Eh, the same could be said about addons but they seem to sell decently.
That is true, I feel most people are happy allowing something to be sold if they are happy with what they get from it and don’t feel the need to make it free to access. It takes effort and motivation to do such a thing.
Some addons provide things which don’t need to be licenced the same as Blender, but modifications to Blender itself inherit the licence. This is likely the biggest difference between many addons and this.
Plenty of addons use GPL as well, so there is no difference in this regard.
I mean 2.79.6 one of the latest builds
Ok I will prepare the file
I’ll to adress all points, feel free to say if I forgot one:
- I already teach everything, there is no secret and all my students have the code.
- the BF accepted some of my patches, some not, the patches in E-Cycles were not submitted thus not rejected either yet.
- All addons on the market are GPL, so the risk is for everyone. But indeed the Blender community seem to be respectful of developers (Python and C/C++ one)
- Red hat, those making Ubuntu, Google with android all make money with GPL code, even in the many million.
- What I offer is not only code, it’s regular builds for 2 platforms on 2 versions, very fast support, new features every month. All reported bugs in my code were fixed under 24 hour.
- Ton wants new cycles dev to be paid by other like for Lukas and Stefan. Working in an open maner was how I started. Trying to make a funding was way too low to live from it, I already said I would prefer to concentrate on code instead of marketing, building, etc.
- Did you ever try to make a review? It’s not as simple as “it’s fast and people are happy and it never breaks so let’s have it in”. Again, there are lot of other factors, like time for review from paid devs, some more obscure most people don’t understand.
Now a question from me: Brecht, Lukas, Sergey, Mai, Stefan Werner all sell their code. What is different with me? That you know the job is already done and they say the job will be done? I find it good they are paid, I find it good I’m paid. And again, I already tried the open bar way, the download/fund ratio was less than 1 of hundred. I think anybody able to buy a render PC can afford 9€/month to get it 2x faster. And in one year, everybody can have it anyway.
The other ways are:
- continue to work for free: Ok if you get me a flat, food, computer, transport, and everything I need for free.
- stop working on Blender, so there is no speedup at all and the 3 people in this thread who think it’s bad making money with Apache/GPL code are happy.
I have the impression many people speak about things they never experienced, basing their reasoning on a beautiful but theoretical world . Learn to code, try to fund your work, make reviews of your code for the BF, then come back here and speak, I think you will have another point of view on the reality.
I had hoped that the “why are you charging money” debate was over. Oh well.
Mathieu’s work is compliant with the GPL. That’s all we need to care about. As long as that’s a given, there is no problem.
Yes, he’s selling Blender builds. Blender builds that contain some code that I wrote in my spare time and contributed for free. So what? I don’t feel cheated, I was well aware of that possibility before I submitted my patches.
If anyone feels like the patches from E-Cycles should be available to everyone for free, go ahead. Buy E-Cycles once, request the source code and publish it. Nobody can stop you. If you want to, provide regular builds for all platforms. Test them, too. Once you’re done with that, you’ll notice that Mathieu is not only contributing code, but that making and publishing builds is work too. Work that he absolutely has the right to get compensated for.
hey. whats about a benchmark for small renderfarms up to 100 slaves?
What do you mean? If you have 100 slaves with cuda rendering, they all will get the speedup. As long as each slave render one frame, it will scale linearly.
We use Blender (cycles) for feature film vfx with mutiple renderfarms. Just want to be sure, that it works in command line rendering before we try it and waste time
See E-Cycles - Faster cuda rendering command line render works.
I’m not debating that he should get compensated for his work, that is clearly his right. I also don’t disagree that he is doing a substantial amount of work which warrants charging money for. For me, personally, the guide on merging patches would be worth the money, considering the hours of time it could save me trying to figure it out on my own.
I’m suggesting that maybe putting potential gains for the entire community behind a paywall is a bit distasteful. Testing changes made to a piece of software is, well, what development is. As for making builds, almost the entire process can be automated, as can be seen with Buildbot.
What would probably look better in the eyes of the community is submitting each change that makes Cycles faster without ruining temporal and spatial consistency to Master, and selling the rest of the (very useful) learning content he is providing. Anything which he has done to modify Cycles (which still hasn’t been described clearly) which isn’t consistent with Cycles philosophy will then be evident and people can more easily determine whether they want those compromises in their version of cycles. Currently it seems like a combination of debatable ‘hacks’ and genuinely useful improvements have been applied, adding up to the “OMG 400% speed increase!!” hype which seems to be driving the majority of community interest.
I don’t doubt he has made some significant improvements to Cycles - which is why I would urge him to submit them to be merged. “The guy who made Cycles 20% faster with a single patch is selling material explaining how he works” sounds a heck of a lot nicer than “Someone is selling Blender with bunch of changes made to it”. I’m not in his shoes so I don’t know what other considerations he has in the decision-making process, I’m sure he has good reasoning for his decisions, but that’s just my thoughts.
You may not be aware of this, but you’re suggesting that he do more work for less pay. Submitting, documenting and polishing a patch until it is accepted to master is extra work on top of developing the actual feature, and can sometimes take up more time than writing the code itself.
Sure, volunteer work can improve your public image. But public image doesn’t pay rent.
I’m not entirely new to developing Blender though I forgive you for thinking so, given that we are on an open forum.
I have made non trivial changes to Cycles and understand that the process of merging a change isn’t simple. I have based my opinions on the knowledge I have and while it is certainly incomplete, I’m not just throwing ideas around with no reasoning.
I understand the implications of my suggestions. That may be the case. It is of my opinion that contributing to open source software is not a viable employment option. I’m a full time developer. I’m sure I’m in the majority in saying that any contributions I might make to Blender occur under the assumption I won’t get any financial gain out of it. Of course that limits the resources I can devote to improving Blender. I agree, public image doesn’t pay rent, but it can lead to a heck of a lot more negotiating power when it comes to finding employment which does pay rent.
That’s how we make our own choices. Mathieu chose to earn his money this way, and I wish him all the best.
Ok, sorry, so then you know a part of my reality at least, yet not all of it. Great if you have enough money and time to afford working for free on a prolonged period of time. I hope I will be able to go back to that state again at some point.
And by the way, the 2.7 version has an option to make the image is 100% like official version. It’s 1.7x faster then instead of 2x in mean, but not a single customer is using it as far as I know, so it proves the other 30% are good for artists.
Those people who decided to use Blender for their professional work maybe can afford a small monthly fee.
The devs aren’t unreasonable, and for those of us who understand what implications certain changes might have, it is understandable that they might not want to merge whatever changes you have made.
I disagree. In certain cases they are. Did you know Cycles in Blender 2.80 Beta and 2.79.6 is 20% slower in render preview? It’s a huge speeddrop and apparently they didn’t care. Look at you own eyes (https://devtalk.blender.org/t/slower-cycles-gpu-in-blender-2-8-beta/4513/5), but here is the point:
“brecht: We’ll try to fix it for the release, it’s not very high priority for me. Progressive refine never had great performance.”
It’s not true at all, Cycles GPU beats V-Ray Next GPU even with the speeddrop. And thanks to Mat it turned out there is a lot more under the hood. I’m really looking forward to what he can do with the render preview and that’s it is worth the money for me.
vray is faaaaaaaaaaaaar better than cycles