hackertube.io

Halide Talk

Andrew Adams · talk · Watch on YouTube

Compilers, interpreters & program analysisSoftware performance & optimizationComputer graphics & rendering

What it is

Andrew Adams, a Halide developer, presents the Halide language for image processing and tensor computation. The talk explains the separation of algorithm from schedule with illustrated examples of parallelism, locality and redundancy tradeoffs.

Why Hacker News recommends it

Recommended as the clearest explanation of Halide's ideas and of scheduling for performance.

  • 5 people recommended it
  • 4 different months
  • 0 strong recommendations
  • first 2014, last 2021

What HN commenters said

  1. “The best way to understand it is probably to watch the video, which is relatively long but explains things clearly”

    Clear, long explanation of Halide's algorithm/schedule separation.

    modeless, 2014-04 · Recommendation · in “Speed memory access by arranging data to take advantage of CPU caching”

  2. “they seem to open new ground in the area you are prusuing”

    Opens new ground in performance work balancing parallelism, locality, redundancy.

    whos3424, 2018-05 · Recommendation · in “How to write efficient matrix multiplication”

  3. “A really nice video with animations of various scheduling approaches for these types of linear algebra algorithms”

    Animations of scheduling approaches for linear algebra algorithms.

    tehsauce, 2018-05 · Recommendation · in “How to write efficient matrix multiplication”

  4. “Here is a relevant video from the creator. It has some fun illustrations”

    Talk from the creator with illustrations explaining how it works.

    all2, 2018-12 · Recommendation · in “Halide: a language for fast, portable computation on images and tensors”

  5. “I feel like this talk on Halide[0] is really illustrative of the general concepts.”

    Illustrates locality and vectorization concepts behind column-oriented layouts.

    T-R, 2021-05 · Recommendation · in “Apache Arrow 4.0”

Excerpts from Hacker News comments, linked to the originals. Summaries and labels were written by a language model from those comments.

Topics: halide, scheduling, performance optimization, locality, vectorization

Also recommended by the same people