danlark

7 ms·

1.
by danlark 2y ago
3 and 4 elements will fail to prove the complexity is linear You still can do 3 or 4 but with slight modifications https://arxiv.org/abs/1409.3600 For example, for 4 elements, it's advised to take lower median for
2.
by danlark 2y ago
Around 4 years ago I compared lots of different median algorithms and the article turned out to be much longer than I anticipated :) https://danlark.org/2020/11/11/miniselect-practical-and-gene...
3.
Miniselect: Practical Linear Time Median Algorithms(danlark.org)
2 points by danlark 2y ago|0 comments
4.
RE2 maintainer Paul Wankadia died(github.com)
70 points by danlark 2y ago|11 comments
5.
by danlark 3y ago
Disclaimer: the author of the comment is the CEO of ClickHouse
6.
by danlark 3y ago
I agree on the sorting front. Removing one cmov is not likely to improve much.
7.
by danlark 3y ago
In this particular case they were universal but in paper it's said the optimizations were done on x86. One of the ideas was to use LLVM IR but intuition for optimizer over optimizer was unlikely to work properly.
8.
by danlark 3y ago
You can see hashing optimizations as well https://www.deepmind.com/blog/alphadev-discovers-faster-sort... , https://github.com/abseil/abseil-cpp/commit/74eee2aff683cc7d... I was one of th
9.
by danlark 3y ago
I am the author of this trick as well You can read about it in https://community.arm.com/arm-community-blogs/b/infrastructu...
10.
“csinc”, the AArch64 instruction you didn’t know you wanted(danlark.org)
3 points by danlark 3y ago|0 comments
11.
ZSTD 1.5.5 is released with a corruption fix found at Google(github.com)
170 points by danlark 3y ago|23 comments
12.
Google Abseil releases performance tips of the week(abseil.io)
2 points by danlark 4y ago|0 comments
13.
Quadratic strict weak ordering check to catch bugs in sorting(github.com)
34 points by danlark 4y ago|1 comments
14.
Miniselect: Selection Algorithms in real world (2020)(danlark.org)
1 points by danlark 4y ago|0 comments
15.
Bit twiddling with Arm Neon: beating SSE movemasks, counting bits and more(community.arm.com)
96 points by danlark 4y ago|7 comments
16.
by danlark 4y ago
Sorry, it definitely was a little inconsistent as I stored all snippets without proper formatting and I tried at the same time align texts for phones, etc Failed at everything. Will improve
17.
by danlark 4y ago
Hi, the author is here I am working in MapReduce/Data-pipelining/Dataproc efficiency internally and externally. In cloud you can check out Google Cloud Dataflow We are working closely with general efficiency and bare metal teams,
18.
by danlark 4y ago
Thanks, I'll remove the note about "recalling"
19.
by danlark 4y ago
You don't, it's just not disallowed :)
20.
by danlark 6y ago
Hi, I don't agree to the end as most of the options are equally applicable to C language except libcxx and others. ThinLTO, inline threshold -- they also can be turned on for C compilation units
21.
by danlark 7y ago
The developer helped us a lot to improve the speed and now it is almost at the same level as jemalloc for us
22.
by danlark 7y ago
Yes, our program ran at half speed.
23.
by danlark 7y ago
We tried mimalloc in ClickHouse and it is two times slower than jemalloc in our common use case https://github.com/microsoft/mimalloc/issues/11