Live data from Hacker News

Viewing profile — cloudhan

cloudhan

HN member
Joined
Wed, Dec 01, 2021, 8:29 AM UTC
HN karma
28
Public activity
20 items

About cloudhan

No profile information was provided.

Recent public activity

  1. comment
    Comment #43502821

    > The official project language is American English with ISO 8601 dates and metric units. Very good taste for an American.

  2. comment
    Comment #41901956

    NPU folks offen time say > it's not enough time to get new silicon designs specifically for Where blahblah stands for a model architecture that has caused a paradigm shift. When yo…

  3. comment
    Comment #41901869

    OK, I am one of the developers in onnxruntime team. Perviously working on ROCm EP now has been transfered to QNN EP. The following is purely devrant and the opinions are mine. So R…

  4. comment
  5. comment
    Comment #40231574

    This reminds me of Weight Agnostic Neural Networks https://weightagnostic.github.io/

  6. comment
    Comment #40231478

    Very unfriendly. The symbolic library (type of activations) requires a branching at the very core of the kernel. GPU will need to serialized on these operations warp-wise. To optim…

  7. comment
    Comment #39557944

    You run or your code runs, choose one and choose it wisely ;)

  8. comment
    Comment #39271073

    The memory is too small to be useful nowadays.

  9. comment
    Comment #38938992

    At the time of the dev of the EP, the tooling is not that good as current. I have imagined a pipeline that HLSL compiles down to DXIL then go through spirv cross and then target wi…

  10. comment
    Comment #38938858

    Not exactly, both cl and glsl can be aot, but the runtime will be limited to some newer version and then the market coverage will be niche, those vendor are so lazy on updating the…

  11. comment
    Comment #38938287

    circle lang is also very worth to check out.

  12. comment
    Comment #38938280

    Thoes "existing libraries" does not necessary mean stdc++, but some parallel primitive, and are essential to performance portability. For example, cub for scan and reduction, cutla…

  13. comment
    Comment #38938104

    I am long sought after a CUDA or HIP compiler that target SPIR-V or DXIL. So that we can compile all thoes neural network kernels to almost all compute devices. The requirements ar…

  14. comment
    Comment #38570908

    Mixtral-8x7B support --> Support new model https://github.com/stanford-futuredata/megablocks/pull/45

  15. comment
    Comment #38570874

    Might be the training code related with the model https://github.com/mistralai/megablocks-public/tree/pstock/m...

  16. comment
    Comment #38304994

    You might also want to watch on helix, if PR 8675 is merged, then scheme will be the extension langauge, then we get the best of both world. Modal editing and scheme =) PR 8675: ht…

  17. comment
    Comment #36237710

    onnxruntime is written in C++. So it is not misleading =)

  18. comment
    Comment #35566362

    This papar is really interesting, especially so when you scroll to the Reference part ;)

  19. comment
    Comment #29427386

    I think you might be interesting in this interactive papar: https://willcrichton.net/nota/

  20. comment
    Comment #29401942

    Then we need a powerpoint virtual machine[1]. [1] https://www.andrew.cmu.edu/user/twildenh/PowerPointTM/Paper....