Live data from Hacker News

Viewing profile — volta83

volta83

HN member
Joined
Thu, Jan 28, 2021, 2:49 PM UTC
HN karma
1,480
Public activity
441 items

About volta83

No profile information was provided.

Recent public activity

  1. comment
    Comment #29653125

    This might be an error on both sides. They want to study how websites handle GDPR and CCPA, and that's probably what they submitted. The IRB reasoned that "websites are not people"…

  2. story
  3. comment
    Comment #29577275

    A simple way to do this is to: MyNestedMaxHeap(T, Ord) = MaxHeap(MaxHeap(T, Ord), Ord) such that when using MyNestedMaxHeap only one Ord type can be passed.

  4. comment
    Comment #29570090

    > You need some way to enforce that the orderings are the _same_ Just require that the Ord type is the same for all heaps ? MaxHeap(MaxHeap(T, MyOrd), MyOrd) uses the same ordering…

  5. comment
    Comment #29539114

    > If you didn't have canonical implementations, you would have to give up being able to write safe & efficient data-structures like this. No? Just make the type generic on the orde…

  6. comment
    Comment #29528720

    I agree with you, which is why I don't really understand what the point of improving FP64 perf by 4x is, if that is not the bottleneck for many apps. Per node, a 4x MI250X node has…

  7. comment
    Comment #29528706

    Makes sense. Comparing nodes with 2x or 4x MI250X vs 4x or 8x A100-80 it doesn't really seem that there is any speed up at all for memory bound apps.

  8. comment
    Comment #29528679

    Then you must be extremely bad at rust.

  9. comment
    Comment #29507658

    Many in the core team don’t code, don’t even know Rust well, etc.

  10. comment
    Comment #29482253

    I wonder how they manage to keep the FP64 units busy. Seems this is an HPC product, but many HPC apps are memory bound. So to improve FP64 perf by 4 one might need to improve DRAM …

  11. comment
    Comment #29433930

    Ideally it would be a constitutional ammendent, stating that privacy is a right of every citizen, and a human right. Legislation to the right to encryption would then just follow f…

  12. comment
    Comment #29428193

    Its not completely unrelated. Excess mortality is the only fact that we know for sure given that every country counts "COVID deaths" differently. Given that there is a statistical …

  13. comment
    Comment #29428187

    Right, but all architectures can handle many combinations of instructions in 1 cycle, so this is not really a great proxy for that. Same for code size. If the instructions are half…

  14. comment
    Comment #29428183

    8x 2 byte instructions (16 bytes) lead to smaller code than 4x 8 byte instructions (32 bytes). Counting number of instructions isn't really a good metric for that either.

  15. comment
    Comment #29423593

    > DPU SmartNICs, Which other companies beyond NVIDIA sell DPUs ?

  16. comment
    Comment #29423534

    I just showed the data, and summarized that according to the data, people in the EU had better chances. I never claimed that this data is the result of COVID policies. This is some…

  17. comment
    Comment #29423474

    The US didn't act as one entity either.

  18. comment
  19. comment
    Comment #29423364

    Can you change the "shape of the vectors? e.g. 1x16 vs 4x4 to support vectors and matrices?

  20. comment
    Comment #29423238

    > RISC-V is faster.. I find it funny that you make the same pitfall than the author did. Faster on which CPU? The author doesn't measure on any CPU, so here there are dozens of peo…

  21. comment
    Comment #29423179

    The CPU executes the two (or more) dependent instructions "as if" they were one, e.g., in 1 cycle. The CPU has a frontend, which has a decoder, which is the part that "reads" the p…

  22. comment
    Comment #29416684

    Good question. Excess mortality in the USA and the EU in 2020 were ~470k and ~580k deaths. The population of the USA and the EU in Jan 2020 was ~329 and ~447 million. The excess mo…

  23. comment
    Comment #29378411

    Running on soft cores.

  24. comment
    Comment #29360216

    You start with BLAS being a factor 2 off, and then go to PETSc, and are another couples of factors off, and then the actual app the scientist wrote, which many use all of the above…

  25. comment
    Comment #29360170

    I think that right now, there is sufficient motivation to implement enough of this to be able to try it out, and then see how it actually behaves in practice. But Rust works on a r…