Tweetovi

Blokirali ste korisnika/cu @trav_downs

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @trav_downs

  1. Prikvačeni tweet
    11. lip 2019.

    Do you like footnotes? Do you like long, rambling lists? Then, this might be the thing for you: There, I smash my personal footnote count record (20). As for lists? It has lists *within* lists. The only thing left is to write it in LISP.

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    1. velj

    It's not every day that a read-after-write makes a server spend 4% of its CPU cycles in one instruction. Wrote that one up here:

    Poništi
  3. 1. velj
    Poništi
  4. 1. velj

    Talk to me like a baby, because I am a web baby (no "full stacker" here...). BTW, this is satisfy all those people who say "but run it on ARM/AMD", so help me out here.

    Prikaži ovu nit
    Poništi
  5. 1. velj

    Rather, I'd like to the one main image displayed and have arrows or buttons or whatever to select the other plots. I guess I need some Javascript blob to do this, but I don't even know what this thing is *called*, so I can't search it. What is it?

    Prikaži ovu nit
    Poništi
  6. 1. velj

    I'd like use an image of a plot on my static HTML blog. OK, I already know how do this part. Next, I'd like to offer variants of the same plot, but perhaps on different CPUs, or some other coarse grained dimension. I don't want to just barf all those plots onto the page.

    Prikaži ovu nit
    Poništi
  7. 1. velj
    Prikaži ovu nit
    Poništi
  8. 1. velj

    In a blog post with footnotes and comments, should the footnotes precede the comments, or vice-versa?

    Prikaži ovu nit
    Poništi
  9. 1. velj

    Shaping up to be an epic series and the visualizations go beyond. Can't wait for parts 4 through 6.

    Poništi
  10. 31. sij

    I think the use of threads for fine-grained/short duration functions is largely unexplored, but this shows you aren't going to do it effectively by starting a dedicated thread each time!

    Poništi
  11. 29. sij

    The big 4 at O2: The big 4 at O2, march=skylake (avx2 for MSVC): Bookmark them!

    Prikaži ovu nit
    Poništi
  12. 29. sij

    In case you ever find yourself setting up the same Compiler Explorer () layout (compilers, args, etc) over and over, this is the obvious solution that I couldn't see:

    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet
    28. sij
    Odgovor korisniku/ci

    After a few unlikely branches, they also start aggressively compiling for size (more than Os), or move things to the cold section.

    Poništi
  14. 27. sij

    A good look at some of the challenges that arise when you try to do fleet-wide profiling. A worthwhile read!

    Poništi
  15. proslijedio/la je Tweet
    26. sij

    Some performance tricks when using the RoaringBitmap library

    Poništi
  16. proslijedio/la je Tweet
    24. sij

    open question for compiler people: what should be taught in a modern undergrad compiler class? I'm revamping / modernizing Utah's version of this and would appreciate feedback

    Prikaži ovu nit
    Poništi
  17. 22. sij

    ... at least if you ignore pointer-to-members, which usually have non-bytewise-zero representations for null.

    Prikaži ovu nit
    Poništi
  18. 22. sij

    I believe the discussion there applies directly to memcpy and memset, but for memset you have the other issue of representation (e.g., maybe a zero int isn't byte-wise zero on you platform), but that kind of falls into the "not important in practice" bucket ...

    Prikaži ovu nit
    Poništi
  19. 22. sij

    I believe the right answer is "it is not safe, even in practice" - specifically due to the chance that your simple& is actually a base class (or, in C++20, a potentially overlapping data member due to no_unique_address). Details:

    Prikaži ovu nit
    Poništi
  20. 22. sij

    One way you might have that guarantee is if you have an array (or array-like) of T (e.g., a T* that must be an array, std::vector<T> or whatever). In that case you know T is the most derived type, so no subobject shenanigans.

    Prikaži ovu nit
    Poništi
  21. 22. sij

    Note that simple is all of: "trivial", "trivially copyable", "standard layout" and POD. So if people tell you any of those are enough to do memcpy or memset safely into an object, they are wrong. You additionally need to guarantee it is not a base class object.

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·