• 0 Posts
  • 27 Comments
Joined 1 year ago
cake
Cake day: June 11th, 2023

help-circle


  • Yup and I am getting sick of hearing this even on Arch Linux. Like, mofo, you could literally run a snapshot or backup before upgrading, don’t blame us if you’re yoloing your god damn computer. Windows have exactly the same problem too and this is why we have backups. Christ.

    On my Arch Linux Install, I literally have a Pacman Hook that would forcibly run backup and verify the said backup before doing a system-wide update.








  • Yup, been writing a new shader language to replace GLSL and HLSL for Vulkan Compute purposed, but I eventually switch from SPIR-V IR to MLIR and use IREE Compiler which accepts the MLIR and compile it to any of CUDA, ROCm, SPIR-V and so forth.

    A lot of it was because of my unadulterated hatred toward our current Machine Learning Frameworks...

    It’s one of the project that I’ve been working on to outright replace Pytorch/Tensorflow and ban those two framework from my office forever. I got fed up not knowing how much exactly do I need in memory allocation, computational cost, and so forth when running or training neural net models. Plus I want an easier way to split the model across lower-end GPU too that doesn’t rely on Nvidia-only GPU for CUDA code. I also wanted to have SPIR-V as a fallback compute kernel, because if CUDA/ROCm is too new for GPU, you’re SOL, but if you have SPIR-V, chances are, any GPU made in the last 10 years that have a Vulkan Driver, would likely be supported.

    One of the biggest plus with MLIR is that you are also future proofing your code, because that code could feasibly be recompiled for new devices like Neural Net accelerator cards, ASIC, FPGA, and so forth.



  • I agree on avoiding on the idea of avoiding having to make your own parser generator, this is precisely what I’m doing and it’s hell. I assumed that you probably want to pick up some understanding on how parser differs when it come to writing grammars. As for ease of use and requiring the least understanding, using something like Earley parser is probably the easiest, it would be slower than other parser algorithms, but it could handle ambiguous grammars making it ideal for first timers to learn how to write a programming language.


  • Yep, and if open source licensing could be revoked on a whim, you can imagine the chaos that ensued. That would be my understanding as well, old version that have MPL license is perfectly fine to fork off, newer version might not be as it is under a different license. One of the reason why I liked Apache License is that it have make it explicitly clear that it’s irrevocable whereas MPL it is operating on an assumption that it’s not revocable. The most fundamental problem with the legal system in USA is that no law is “set in stone” and leaving things to assumption is open to reinterpretation by the judge who may have sided against you. (Hell, Google vs Oracle on Copyrighted API is still on case-to-case basis, so take it as you will.)

    Disclaimer: I am not a lawyer. I just share what I learned from Legal Eagle youtube and few other sources.



  • I concur, there was a few problems that might come up on various platforms like Windows not implementing C11 standard threads and other stuff, you would instead use TinyCThread library that works like a polyfill.

    All problems and challenges are workable, if the problem with Debian is out of date library, you could set up CI/CD for release build that rebuild your software when update occurs and static link the updated dependencies.

    Back to your point, if they didn’t design their code and architecture to be multiplatform like in C, they need to re-evaluate their design decisions.






  • It just rooted back to my frustration when I was trying to fill in missing implementation details on projects like Skia (at the time it lacked support for Vulkan.) My very fundamental core belief is that for core libraries like say, Skia, Neural Net Framework, and other crucial projects like that should offer a way in C API that allows every type and implementation to be extended upon by any other language that can interface with C API by providing your own VTable or whatnot.

    One of the approach I do for my GUI Toolkit written in C (specifically on Linux to replace QT and GTK) was making a single inheritance object oriented programming in C.and if you insert the base class type structure at the top of your custom struct type and provide your own VTable for those objects, you can readily extend the underlying library natively in whatever programming language you use assuming it can talks to C API in a complete sense.

    Let me know if you want a demonstration of this, I would be happy to find the time to set up a small sample to give you the idea on how it’s done.

    And I am also aware of the criticisms on those approach, verbosity of attempting to implement object oriented programming in C is kind of absurd and the API coverage would balloon. That is largely why I work on a Compiler-Generator Framework specifically to address the challenges by allowing me to add dialects on top of C Language such as generic, object oriented programming, and various dialects. I brought C closer to C# in term of syntax and features and at the end of compilation, it still produces readable C language code output and it also generates what I called an FFI-JSON. It’s essentially a JSON file that describes all of the types used in a C project, the sizes of integers/floating points, structure types and it’s fields/offsets/sizes comments, and function declarations. It’s done in a way that you could read the JSON file and generate your programming language binding library saving you weeks of work.