wmu

11 ms·

1. 61.▲
   by wmu 8y ago

   Your approach is incredibly fast. I love the code, it's so short. :) Please take look at https://github.com/WojciechMula/toys/tree/master/avx512-remo... , I put there some raw numbers; will update th
2. 62.▲
   by wmu 8y ago

   > I've been nerdsniped as well. I can't say I'm going to go ahead and try and solve it, but the methodology presented in the post seems suboptimal. Let me explain it. I do know the presented approach is extremely naive, b
3. 63.▲
   by wmu 8y ago

   This is a problem with base64 decoding. While there are vectorized algorithms for decoding and encoding, they do expect a text without spaces or newlines. However, by design, in emails the base64-encoded data must be split to lines not exce
4. 64.▲
   by wmu 8y ago

   I was wondering if you like to contribute better scalar code? I'll be happy to include your (or anybody else) code and then compare different approaches.
5. 65.▲
   by wmu 8y ago

   TBH I didn't think about throttling. My (mis)understanding is that a single program which completes in a fraction of second on an almost-idle machine doesn't suffer from AVX512 throttling. Daniel has reviewed this problem in serie
6. 66.▲
   by wmu 8y ago

   Hi, author here. That's amazing! I totally love this approach, it so neat. Wow :) Would you mind if I include your solution in my repository and the article? BTW you can always check validity of your code with Intel Software Developer
7. 67.▲
   by wmu 8y ago

   Any examples?
8. 68.▲
   by wmu 8y ago

   I agree about open source projects, then compiler options should not be so strict. But during development it might be useful; I use -Werror before releasing piece of code, just not to add any new warnings. BTW, a few years ago we had a nast
9. 69.▲
   by wmu 8y ago

   "-Wall -Wextra -Werror" are you allies :)
10. 70.▲
    by wmu 8y ago

    AdBlock might help.
11. 71.▲
    by wmu 8y ago

    Bruce's write-ups are just great, sheer joy of reading.
12. 72.▲
    by wmu 8y ago

    Are there any benchmarks that show how fast this db is?
13. 73.▲
    by wmu 8y ago

    Last activity on github was 3 years ago. Seems the project is dead.
14. 74.▲
    by wmu 8y ago

    Sorry, I forgot that in HN comments the asterisk char is an italics indicator. There should be a mark after SSSE3.
15. 75.▲
    by wmu 8y ago

    Speaking of the Intel world it's not that bad. There are three major version right now: SSE4.1, AVX and AVX2 (AVX512 is not popular yet). In the past (roughly 10 years ego) it was a problem, as there were: MMX, SSE, SSE2, SSE3, SSSE3 ,
16. 76.▲
    by wmu 8y ago

    The title might look like a clickbait, but the article is interesting and well written. :)
17. 77.▲
    by wmu 8y ago

    Really nice overview.
18. 78.▲
    by wmu 8y ago

    "A trie is basically a special case of a DFA." The best definition of trie I read.
19. 79.▲
    by wmu 8y ago

    My fault, I just skimmed through the article and missed that comparison.
20. 80.▲
    by wmu 8y ago

    Comparing lookup time of unsorted vector and sorted set is unfair. The vector should be sorted and then we might compare plain binary search against BST performance.
21. 81.▲
    by wmu 8y ago

    Speaking of using non-ASCII and non-English names in program, I'd suggest not to use them. People who don't know the alphabet/language will understand literally nothing; it's much better to use English names.
22. 82.▲
    by wmu 8y ago

    > Register renaming, cache hierarchies, out of order and speculative execution etc are not visible at the assembly / machine code level Cache hierarchies are directly accessible with CLFLUSH, INV, WBINVD x86 instructions; we may cou
23. 83.▲
    by wmu 8y ago

    Sorry, have no idea.
24. 84.▲
    by wmu 8y ago

    Author here. This was true several generations ago (core2, for instance), now the performance penalty is negligible.
25. 85.▲
    by wmu 9y ago

    Here is the link to the blogpost that describes the problem: http://www.dkriesel.com/en/blog/2013/0802_xerox-workcentres_... ?
26. 86.▲
    by wmu 9y ago

    Better would be the Huffman coding; one has to deal with variable bitfield widths.
27. 87.▲
    by wmu 9y ago

    Thanks for this clip, it's amazing how it keeps the head still. BTW, the kingfisher has similar ability, check this https://www.youtube.com/watch?v=NIs4tpV2FUQ
28. 88.▲
    by wmu 9y ago

    There were similar stories from the warehouse located in Poland.
29. 89.▲
    by wmu 9y ago

    I don't understand one things: the original scalar version sums bytes, while the SIMD version sums 32-bit values and returns the sum mod 256. The SIMD version might be much simpler if just byte-wide operations were used. Did I miss som
30. 90.▲
    by wmu 9y ago

    I'm a low-level programmer and it's the first time I come across the linear-feedback shift register. I would say it's rather a part of digital circuits design. Now, when FPGAs become more and more popular, people will be forc