I create rust based version of klipper

Developers
1

i create rust based version of klipper with help of ai assistant

i need some code review and on hardware test

2

I don’t know enough to help with code review and testing. Perhaps there is a suite of tests used for klipper regression testing? What is your end goal?

3

my propose is make it faster and lighter to eventually run one single micro controller like esp32-s3

4

Do I get you right? You want to use an esp32-s3 as a Klipper host?

5

Faster? For what? You can’t exceed the limit of the mechanics anyway.

6

I guess, Faez Barghasa thinks a rust https://rust-lang.org/ version of Klipper is faster in code execution than the official Klipper to use an esp32-s3 as a Klipper host. I doubt he will succeed.

@FaezBarghasa I never asked about your incentive for the project. I’m sure you spend some time.

1 Like
7

Interesting. Did you benchmark it?

I am curious if it can beat the step rate or command dispatch.

Also, it can be a little complicated to do async stuff and make it compatible with the current behavior for I2C/SPI sensors, I guess.

Thanks.

8

I’ve been using Sonnet 4.5 on my klipper project. It does best at things where there are a lot of examples and everything is written down. It has blind spots when you start integrating with hardware because part of the system is hidden from it. ChatGPT will casually rewrite your hard-won register read sequence because it can’t see the point of what you’re doing (breaking it in the process). I am interested in seeing how this project turns out. It will be a good learning experience if nothing else.

9

I’ve been thinking about this a bit more….

The klipper ecosystem is an amazing resource. So many capabilities ready to use without development, but the key ingredient is the community of people associated with it.

One aspect of the community that is wisdom is the decision not to spend resources trying to debug custom development issues. Custom development divides the resources, do that enough and you lose focus and stop making progress, and then the community begins to shrink.

Big picture decisions need to be made with the community as the highest priority. I feel that a rust based port is an interesting challenge and test of AI. Let’s say it is a complete success, what’s next? Host processors are reasonably priced, so what problem is it solving? I feel like the existing solution is “good enough” for everything but niche applications.

All that to say, any path that creates divides in the community seems counterproductive. If the community decides to pursue a rust based solution it has a chance, otherwise it’s just a cool trick.

1 Like
10

Klipper is in my eyes a one person driven project. Named. Kevin O’Connor or Kevin or here koconnor. I guess he is a very private person. The only thing I could find in the Internet regarding koconnor and Klipper is this https://www.youtube.com/watch?v=tODfTn9Yr8s, pretty interesting.

Just my 2ct: When I started with Klipper, I was very, very skeptical that this construct/concept could succeed. But…amazing.

I don’t really understand, what you want to explain. I think this community is very much involved in the progression of Klipper. For example, you can’t just start a pull request on github. Your pull request will be first discussed here in the community and than maybe okayed.
2nd: Everybody gets help here and we learn from the feedback, when questions are asked (at least my impression).

Answered above.

Regarding rust. Are you aware how precisely time driven Klipper is (please community correct me, if my statement is nonsense)?
How do you want to make sure the identical time precision using rust?

Regarding AI, I’m absolutely open for AI for Klipper. I love AI for machines. All the wonderful people in this yet wonderful world will get more stupid, when using AI to search the Internet. I’m fu**ing afraid!

Please explain more detailed. I don’t get your point. Maybe somebody else here can help me out here?

I don’t understand the meaning of that sentence. What do you want to express?

edit:
Why do you want to reinvent a very well running concept to rust? In the economical world you just change concepts (hardware and/or software) to increase your margin.

edit: I forgot.
@FaezBarghasa
What about

1 Like
11

@hcet14 I’m new here, so perhaps I have spoken out of turn. I’ve certainly failed to communicate. Neither was intentional, my apologies.

We agree that klipper is something of value. It took focused effort to create. I don’t see much of a return on investment for attempting to port klipper to Rust. It’s a worthy challenge for an individual, but the results are not likely to be generally useful.

12

i am embedded rust developer and it doesn’t take my time as you consider

as you can see in mu documents , i used mks skipr , and i had not good experience as user , host app seems laggy and not smooth enough fo me.

my last board was mks tinybee and it doesn’t support klipper , so i think about rewriting whole project in rust , that klipper could use rusts concurrency features for speed and borrow checker features for memory optimization so it can had better user experience and cheaper hardware can use klipper performance.

13

I think it is a worthy challenge and am interested in your progress and results.