Live data from Hacker News

Viewing profile — superdimwit

superdimwit

HN member
Joined
Mon, Apr 09, 2018, 4:08 PM UTC
HN karma
210
Public activity
76 items

About superdimwit

No profile information was provided.

Recent public activity

  1. comment
  2. comment
    Comment #41291201

    Clang recently got a new calling convention that makes these tail calls much cheaper (avoids the need for the caller to preserve some registers). I can never remember the name - it…

  3. comment
    Comment #36383280

    https://eli.thegreenplace.net/2012/07/12/computed-goto-for-e... is a good summary

  4. comment
    Comment #36383265

    On 1: this small amount of overhead matters because the amount of work you do on each opcode can be tiny. The extra jump could be 20% of your runtime! On 2: yes, this helps the ind…

  5. comment
    Comment #35725955

    Try dear-imgui + implot. There are Python bindings

  6. story
  7. comment
    Comment #33383296

    There’s a Hello World on the page

  8. comment
    Comment #32971888

    30 years?

  9. comment
    Comment #32971844

    Agreed, autovec is in my experience where clang really shines.

  10. comment
    Comment #31897413

    Disappointed this wasn't an article about bath houses

  11. comment
    Comment #31639338

    One nice property is their stable addressing. A linked list element won't move once allocated.

  12. comment
    Comment #30588734

    Amsterdam is good for options market making, most of the techniques in this book come up a lot in trading

  13. comment
    Comment #30036040

    reinterpret_cast comes up a lot when handling binary protocols - e.g. reading a message struct off a buffer / packet. How would this be handled in Rust or otherwise without copying…

  14. comment
    Comment #27905613

    You can just define a generic fallback method like `eff(p, d1, d2) = eff(p, d2, d1)`

  15. comment
    Comment #27813063

    Could a difference in alignment of the hot loop also have an effect here?

  16. comment
    Comment #27785313

    QR, Cholesky

  17. comment
    Comment #27575633

    I most often need to restart the REPL when redefining structs.

  18. comment
    Comment #27574504

    If they don't, and can't demonstrate that they're doing something else useful with that money, investors get angry and fire the board.

  19. comment
    Comment #27281808

    I am hopeful that Julia should be able to get this cross-session caching of compiled code. Would make restarting the REPL (to e.g. add a field to a struct) much less frustrating.

  20. comment
    Comment #27185220

    even better, don't leave it as a comment. Leave it uncommented, as a syntax error, so when you try and run your code on Monday morning you are reminded of it!

  21. comment
    Comment #27071916

    Big if/else lists will often get compiled into a jump table anyway.

  22. comment
    Comment #27043269

    Go Channels?

  23. comment
    Comment #26767369

    ARKK is far from passive

  24. comment
    Comment #26584354

    A low-latency GC would also be great. But again, the JVM only has that due to many millions of dollars spent over decades.

  25. comment
    Comment #26584338

    In your experience, what are the current limitations?