Skip to content
HomeVideos
AllLongsCategoriesPlaylists
BooksBlogsPodcasts
HistoryListsWatch laterLiked videosFollowing

The videos, books, blogs and podcasts Hacker News keeps recommending, all in one place.

VideosBooksBlogsPodcastsAboutPrivacySettings

Data from Hacker News, HackerTube, HackerBooks, HackerBlogs and Open Library. Credits · Not affiliated with Y Combinator.

HomeVideosBooksBlogsPodcasts
Ctrl K

    Stanford Seminar: Beyond Floating Point: Next Generation Computer Arithmetic

    Stanford Online

    5 HN recommenders · recommended in 9 months, 2017–2021

    YouTube

    1:31:28 · 33K views · Feb 3, 2017

    John Gustafson presents posits, a proposed drop-in replacement for IEEE 754 floats that is simpler in hardware and more flexible in precision. The seminar includes a live demo and machine learning results.

    Cited in floating-point threads as the main introduction to posits and unums as an alternative to IEEE floats.

    Hardware & ElectronicsComputer architecture & CPUsPure mathematicsMachine learning & deep learning
    1. 1:10:03
      Stanford Seminar - Generalized Reversible Computing and the Unconventional Computing LandscapeStanford Online16K views • 9 years ago
    2. 1:30:50
      Lecture 7: Introduction to Galois Fields for the AES by Christof PaarIntroduction to Cryptography by Christof Paar278K views • 12 years ago
    3. 44:09
      Lec 1 | MIT 6.042J Mathematics for Computer Science, Fall 2010MIT OpenCourseWare2.8M views • 13 years ago
    4. 1:34:42
      The Great Debate @ARITH23: John Gustafson and William Kahandancefan4.2K views • 10 years ago
    5. 9:16
      Floating Point Numbers - ComputerphileComputerphile2.5M views • 12 years ago
    6. 1:08:40
      Lecture 1 | Machine Learning (Stanford)Stanford2.7M views • 18 years ago
    7. 1:14:48
      The Future of Mathematics?Microsoft Research116K views • 7 years ago
    8. 1:30:45
      The Simple Essence of Automatic Differentiation - Conal ElliottMicrosoft Research17K views • 8 years ago
    9. 1:10:58
      Trends in Deep Learning Hardware: Bill Dally (NVIDIA)Paul G. Allen School32K views • 2 years ago
    10. 1:21:28
      Lecture 01 - The Learning Problemcaltech1.5M views • 14 years ago
    11. 55:02
      GradIEEEnt half decent: The hidden power of imprecise linessuckerpinch391K views • 3 years ago
    12. 1:12:56
      Lecture 1A: Overview and Introduction to LispMIT OpenCourseWare421K views • 7 years ago
    13. 1:29:55
      Sorting Algorithms: Speed Is Found In The Minds of People - Andrei Alexandrescu - CppCon 2019CppCon184K views • 7 years ago
    14. 2:25:52
      The spelled-out intro to neural networks and backpropagation: building microgradAndrej Karpathy4.4M views • 4 years ago
    15. 1:20:33
      Lecture 1 | Convex Optimization I (Stanford)Stanford754K views • 18 years ago
    16. 40:13
      "Programming a 144-computer chip to minimize power" - Chuck Moore (2013)Strange Loop Conference19K views • 5 years ago
    17. 59:23
      28c3: The Science of Insecurity28c342K views • 14 years ago
    18. 1:00:43
      Jim Simons (full length interview) - NumberphileNumberphile21.4M views • 11 years ago
    19. 20:08
      Fast Inverse Square Root — A Quake III AlgorithmNemean6M views • 5 years ago
    20. 1:07:22
      Invited Talk - Guy SteeleClojureTV39K views • 9 years ago

    13 comments from Hacker News

    1. @trsohmersFeb 2017· in “The “high-level CPU” challenge (2008)”

      Well unums in particular (especially the new type 3 "posits" ones) are very interesting and practical for hardware implementation. John Gustafson, the creator of unums, gave the first talk on type 3 unums at Stanford earlier this month, and there are already multiple hardware implementation efforts (one of which I am indirectly overseeing). While many people have recently hyped up low precision floating point for things like machine learning, Posits provide greater dynamic range AND greater accuracy for very few bits (as low as four, but with good ML starting 8 to 10 bits). On top of all of that, it costs less area and power in silicon than IEEE hardware. Check out the talk: this video

      RecommendsView on HN
    2. @vanderZwanFeb 2017· in “GCC optimized code gives strange floating point results”

      I'm very surprised nobody has brought up John Gustafson's Unum work. His latest iteration is the posit, which he presented at Stanford earlier this month[0][1] EDIT: apologies to CalChris[2]. What I should have said is discussed Gustafson's Unum work. [0] this video (relevant question at 1h22m37s) [1] [LINK] [2] [LINK]

      RecommendsView on HN
    3. @dnauticsSep 2017· in “Floating Point Visually Explained”

      You may like: this video It turns out the math is slightly harder but it's faster than IEEE FP in hardware, probably because of less conditionals in the spec. There is better performance in terms of numerical accuracy at the cost of substantially more difficult error analysis. Disclaimer: I'm in the video.

      RecommendsView on HN
    4. @gbrown_Jun 2018· in “Floating-Point Determinism (2013)”

      For anyone interested in this sort of thing you'd probably also be interested in John Gustafson's work on unums/posits. Here's a nice introductory presentation this video

      RecommendsView on HN
    5. @espeedFeb 2017· in “Why does e to pi i equal -1? (2015) [video]”

      Have you seen "Stanford Seminar: Beyond Floating Point: Next Generation Computer Arithmetic"? this video

      RecommendsView on HN