📜 ⬆️ ⬇️

Torrent render for 3ds max - almost everything will have to be redone

image

I continue my series of articles on the torrender.net project, since the current situation is again far from my previous plans, and all the most interesting for me, perhaps, will begin in the near future.

Previous articles:

The history of the creation of a torrent render for 3ds max
History of torrent rendering for 3ds max

Background:

So, four months have passed since the release, during this time I finished the project to the stage when it became impossible to add anything fundamentally important through MAXScript, in particular, I implemented a single-frame render breakdown for different views and collecting pieces into one picture on the lich side through ImageMagick, this feature was not ready at the time of release, and I had high hopes for it, since the max is used mainly for architectural visualizations, and it happens that rendering one frame lasts three and four hours, that's what they want will appear sya and the system will finally become active, since there are many times more people who want to render statics than those who render animation. But I was wrong, the static break feature passed by users, there was no job, or any significant number of new nodes, despite the fact that I posted information about statics on one of the most popular forums “interior design maksistov” - 3ddd.ru, did not help. And, if so, then the thought appeared by itself - and what is the point for me to introduce a local network render, connect custom render farms (this was the next and last development stage according to the plan), even if nobody wants to render the static? And is there any sense in general to continue, if the users of all this do not need it, then what do they generally need?
')
The question is, there is a simple answer - the button "render beautiful"

Having read here and there (and on Habré including) various comments, I came to the following disappointing conclusions for myself:


As a result, the torrent client should be similar to existing applications for local network rendering (such as backburner) and work as a local render farm, only the nodes are located not in the next room, but somewhere on the Internet. We get: any number of nodes on a job, jobs of any size and complexity, splitting a single-frame by different sidas (you can render heavy statics), rating system, friends mode, connecting local farms, and all this is free for the end user. Well, by analogy with torrents: in ordinary torrents, the user downloads the file that he needs and starts distributing it. In the torrent renderer, he distributes the Joba file and downloads frame files (or frame fragments) - distributes one, downloads another.

What will have to do:



Why C #? In fact, I first wanted to go on a simple path and try to do it on AutoHotkey, but then I remembered that I had a lot of dotnet inserts on a max script and in C # it would be much easier with them, and besides, if I choose a new language for studying, it is better to have something larger than AHK.

After thinking about the pros and cons, remembering previous years of developing and picking in PHP, I realized that the idea of ​​redesigning the client was essentially from scratch, plus I didn’t like studying another language at all if, again, everything will be only on my pure enthusiasm and without any guarantee of success of the future system. Here we need to insert a picture where a glob of light turns on in the gloomy genius in my head, because the idea of ​​“Russian crowdfunding” came about like this, immediately materializing into the next post on the center:
Currently, I no longer have the opportunity to continue development only with my own funds, because money and free time tend to run out and therefore the project is going to crowdfunding, I’ll choose a specific platform the other day. The fact that I personally need the project code, I have already done, but give the users completely different features, I suspect, so the nodes are offline. Actually, a fundraising is needed to fulfill user desires (I did my own for free), and if we don’t dial it means no one needs it, I’ll save time and I won’t go on free of charge.

New features will require a lot of time to implement and I don’t want to risk harnessed on such a scale on the same bare enthusiasm:
1. I will rewrite the client from the max script to another development language, it will work in the Windows tray unnoticed and independent of max (as a torrent client)
2. Max is used and started only for creating jobs and for rendering, all client logic that is not directly related to max and renderer will work without max - this is also a reserve for other packages
3. Self-selection of all files accompanying the scene (textures, etc.) by the machine when creating the Joba
4. 100% scene protection - no scenes))) Ie no max files. If anyone remembers, there was such a script from Bobo, who translated the current scene into a set of max script commands, i.e. in fact, it generated a script that creates a similar scene when it is launched — I will do something like this. Without a client, it will be physically impossible to open such a job, and since the client will not be in a max script, hackers will not be able to quickly break it and tear out the parser (if they succeed at all).
5. Scenes and render results can be distributed not only by torrent, but also uploaded to ftp or sent via email.

ps If the crowdfunding is successful, then the project will have to be rebranding, there will be a new name and a new domain.

Well, and then the matter of technology, firmly deciding to start crowdfunding, I actually launched it on planeta.ru, but for a month the project managed to collect only 7 thousand out of 200 thousand rubles, less than a week left, so I won’t be surprised if this thing will not burn out, and again I will have to puff on enthusiasm in my free time, as always, selyavi.

Source: https://habr.com/ru/post/228225/


All Articles