Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
ot
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
48 ms
·
541.
▲
by
ot
9y ago
You don't even have to go into SIMD and lookup tables: on 64-bit archs you can process 8 bytes at a time using a few bit tricks. Lookup tables work great in microbenchmarks because the tables are always in cache, but not so great if th
542.
▲
Real-Time Adaptive Image Compression
(arxiv.org)
1 points
by
ot
9y ago
|
0 comments
543.
▲
Core-local Statistics
(rocksdb.org)
2 points
by
ot
9y ago
|
0 comments
544.
▲
by
ot
9y ago
@luu you submitted two links about dynamic succinct data structures today, what picked your interest?
545.
▲
OSS-Fuzz – continuous fuzzing of open source software
(github.com)
2 points
by
ot
9y ago
|
0 comments
546.
▲
RocksDB: The birth of RocksDB-Cloud
(rocksdb.blogspot.com)
2 points
by
ot
9y ago
|
0 comments
547.
▲
Announcing Fairseq
(facebook.com)
2 points
by
ot
9y ago
|
0 comments
548.
▲
by
ot
9y ago
Python's regex library compiles the regex to a sequence of opcodes and then does recursive descent, so yes, it's exponential.
549.
▲
Deep Learning for Program Synthesis
(microsoft.com)
1 points
by
ot
9y ago
|
0 comments
550.
▲
by
ot
10y ago
That was my point: I'm not saying that phones will replace cameras for all uses (there are inevitable physical limits), just that cameras will look very different from what they look now.
551.
▲
by
ot
10y ago
Yeah but I think that the timeframe for smart cameras is a lot closer. Much of the theory behind computational photography has been well understood for a while and it's basically physical modeling and linear algebra. It just requires l
552.
▲
by
ot
10y ago
The rise of computational photography is going to considerably change the camera landscape. This is already getting mainstream with the "depth effect" in the iPhone 7+ and I believe that Google phones are doing something like that
553.
▲
Designing 100G optical connections
(code.facebook.com)
95 points
by
ot
10y ago
|
46 comments
554.
▲
by
ot
10y ago
Unless you're talking about extremely long strings, computing a hash takes far less time than a cache miss.
555.
▲
by
ot
10y ago
If you don't need sorted search (lower/upper bound) or prefix search I doubt even a good trie would beat a good hash table. Have you tried google::dense_hash_map? std::unordered_map is slower because it uses chaining to resolve co
556.
▲
by
ot
10y ago
Wouldn't they also run into the licensing problem, even more so if they try to sell access?
557.
▲
by
ot
10y ago
That's not UB. Argument evaluation order is unspecified , not undefined.
558.
▲
Happy 30th birthday to GCC
(shape-of-code.coding-guidelines.com)
19 points
by
ot
10y ago
|
2 comments
559.
▲
Errata Security: Assert() in the hands of bad coders
(blog.erratasec.com)
4 points
by
ot
10y ago
|
0 comments
560.
▲
by
ot
10y ago
Thanks for saying this! It is incredible how widespread still is the belief that FNV is a good hash function, when much better ones exist that are faster even on short keys.
561.
▲
by
ot
10y ago
> A perfect hash function is one that maps a finite set of keys into the same number of buckets with no collisions. To be even more precise, that's a minimal perfect hash function. "Perfect" only means that it produces n
562.
▲
Amazon Prime Air drone completes its first US public delivery
(cnet.com)
3 points
by
ot
10y ago
|
0 comments
563.
▲
Triple Pendulum CHAOS
(jakevdp.github.io)
1 points
by
ot
10y ago
|
0 comments
564.
▲
by
ot
10y ago
In case anyone else is wondering, this is the original Slashdot discussion: https://ask.slashdot.org/comments.pl?sid=10352645&cid=540224... I don't think it was linked in the post.
565.
▲
Automatic Smoothing for Attention Prioritization in Time Series
(futuredata.stanford.edu)
3 points
by
ot
10y ago
|
0 comments
566.
▲
by
ot
10y ago
I might have given the wrong impression, see my other comment below. "Clowntown" is almost always directed at ourselves, in a self-mocking, playful way. Also see answers to this Quora question: https://www.quora.com
567.
▲
by
ot
10y ago
That word is more often than not directed at our own work. Maintaining a self-mocking attitude can help in being more receptive to valid criticism.
568.
▲
by
ot
10y ago
I see. I didn't know "amateurish" had such negative connotations. Unfortunately HN doesn't let me edit the post anymore.
569.
▲
by
ot
10y ago
If you read carefully you'll see that "clowntown" referred to GCC's implementation of std::hash, not the post. I think it's natural to have high standards for GCC. I don't think my tone was too aggressive, mayb
570.
▲
by
ot
10y ago
> Python uses the identity function for number hashes, because if you're inserting mostly sequential numbers into a dict this gives you better cache performance. That's a good idea when you control both the hash function and th
More ›