home.social

#swar — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #swar, aggregated by home.social.

  1. Now I see that #vibecoding is a good educational tool if you are lazy but still get pleasure from using your brain.

    The first version was a slow naive implementation. An LLM optimized it with #bitwise tricks, but was bad at explaining them. Another LLM explained the algorithm with a guide on Single Instruction Multiple Data (within-a-register) processing.

    I've never encountered these tricks in my working routine and probably would have never tried them without a practical need. Now I'm at least aware of them, thanks to playing with evil ai. Everything at zero cost.

    #SIMD #SWAR #ConnectIQ

  2. Now I see that #vibecoding is a good educational tool if you are lazy but still get pleasure from using your brain.

    The first version was a slow naive implementation. An LLM optimized it with #bitwise tricks, but was bad at explaining them. Another LLM explained the algorithm with a guide on Single Instruction Multiple Data (within-a-register) processing.

    I've never encountered these tricks in my working routine and probably would have never tried them without a practical need. Now I'm at least aware of them, thanks to playing with evil ai. Everything at zero cost.

    #SIMD #SWAR #ConnectIQ

  3. Hey all! 👋🏻
    I’m looking for some shader-like pipeline/#rendering system/library/framework for 1-bit graphics with 2x #framebuffer (double-buffered — actual & previous) with #blitting on #SIMD and #SWAR? CPU-only, mostly targeting ARM32/64/Thumb1.
    I understand that it’s rare and mostly impossible to exist, so I just need some source-based guidance/hints of oldschool/demoscene- tricks and algorithms which I don’t know yet (I know a lot already, I’m 40)) and of course i can port.

  4. Hey all! 👋🏻
    I’m looking for some shader-like pipeline/#rendering system/library/framework for 1-bit graphics with 2x #framebuffer (double-buffered — actual & previous) with #blitting on #SIMD and #SWAR? CPU-only, mostly targeting ARM32/64/Thumb1.
    I understand that it’s rare and mostly impossible to exist, so I just need some source-based guidance/hints of oldschool/demoscene- tricks and algorithms which I don’t know yet (I know a lot already, I’m 40)) and of course i can port.

  5. Hey all! 👋🏻
    I’m looking for some shader-like pipeline/#rendering system/library/framework for 1-bit graphics with 2x #framebuffer (double-buffered — actual & previous) with #blitting on #SIMD and #SWAR? CPU-only, mostly targeting ARM32/64/Thumb1.
    I understand that it’s rare and mostly impossible to exist, so I just need some source-based guidance/hints of oldschool/demoscene- tricks and algorithms which I don’t know yet (I know a lot already, I’m 40)) and of course i can port.

  6. Hey all! 👋🏻
    I’m looking for some shader-like pipeline/#rendering system/library/framework for 1-bit graphics with 2x #framebuffer (double-buffered — actual & previous) with #blitting on #SIMD and #SWAR? CPU-only, mostly targeting ARM32/64/Thumb1.
    I understand that it’s rare and mostly impossible to exist, so I just need some source-based guidance/hints of oldschool/demoscene- tricks and algorithms which I don’t know yet (I know a lot already, I’m 40)) and of course i can port.

  7. Hey all! 👋🏻
    I’m looking for some shader-like pipeline/#rendering system/library/framework for 1-bit graphics with 2x #framebuffer (double-buffered — actual & previous) with #blitting on #SIMD and #SWAR? CPU-only, mostly targeting ARM32/64/Thumb1.
    I understand that it’s rare and mostly impossible to exist, so I just need some source-based guidance/hints of oldschool/demoscene- tricks and algorithms which I don’t know yet (I know a lot already, I’m 40)) and of course i can port.