gorset

9 ms·

31.
by gorset 8y ago
Thanks for the quick follow-up! This might be a difference in expectations, where we are talking about two styles of merge requests. I believe they can be both supported well by gitlab (in fact, I consider gitlab to be the best review tool
32.
by gorset 8y ago
I feel they're making it harder to review a merge request with multiple commits. It used to be able to click on a commit and get succinct view of the actual commit with title and message. Now days after clicking on a commit, you have t
33.
by gorset 10y ago
We're using grpc-java in production for some of our based backend system, slowly replacing our old netty/jackson based system using JSON over HTTP/1.1. The performance is good, and it's nice to have proto files with mess
34.
by gorset 14y ago
According to the compute engine talks from this year's google IO, they have their own private network covering most of the world and uses anycast for external Internet Addresses. This basically means that a connection from you will be route
35.
by gorset 14y ago
I think he risks dropped connections because the listen queue can overflow. He doesn't mention increasing somaxconn, which means it probably has a default of 128 (I don't think increasing tcp_max_syn_backlog is needed since he's using synco
36.
by gorset 14y ago
This is premature optimization. Do you think it's expensive to write a few brackets, quotes and escape a few characters? We have harder problems to think about than optimizing for writes in our logging system :-) How can you trade write/rea
37.
by gorset 14y ago
Remember we're talking about writing to disk. The time it takes to marshal the data is pretty much insignificant to the time it takes to do the actual write. We use JSON on disk heavily in several places where we generate or receive huge am
38.
by gorset 15y ago
Performing the intersection of sorted lists is a well studied problem in information retrieval, so nothing new or surprising here... The performance can be improved even further by for example using a compressed integer list with an embedde
39.
by gorset 15y ago
> a) How can you make standard-OOP-like method dispatch faster than vtables? The JVM optimizes this by using (P)ICs, which are faster than regular vtables. http://en.wikipedia.org/wiki/Inline_caching
40.
by gorset 15y ago
My solution is to use a fixed tree with bloated leaf nodes (use more bits per element in the leaf node). A high number of nodes with a low number of elements causes a high overhead, so I simply rebuild the tree at a few given thresholds to
41.
by gorset 15y ago
http://erik.gorset.no/ not much content, though..
42.
by gorset 15y ago
Cool, WT and succinct data structures are definitely an interesting field these days. I'm primarily using as a compact order preserving structure to support fast lookups for both (token -> position) and (position -> token). I'm usin
43.
by gorset 15y ago
You also get order with a wavelet tree. You can ask "Give me the absolute position for the 4th value 42 in the array". By using a wavelet tree with an index, you can answer that question in almost constant time.
44.
by gorset 15y ago
Wavelet trees are awesome. There are several ways to speed up the rank/select operations, which the article doesn't mention. For example, you can create a simple (popcount, offset) index for each bitmap. Then when you need to find the nth b
45.
by gorset 15y ago
This is a myth. Or more precisely, a mmaped region no longer in use will be unmapped whenever the JVM feels like it (which can, of course, be never). Try running the loop: long size = 0; while (size >= 0) { size += Files.map(v
46.
by gorset 15y ago
Are you a human or a bot? The video is a 4m47s lightning talk which is basically a TL;DW for a real talk.
47.
by gorset 15y ago
From the architecture overview; they need 7 GB of memory to index 100 million objects. That's over 75 bytes per object, which is quite a lot. To be able to guarantee maximum 1 disk access, they would need to store the hash, the address, and
48.
by gorset 15y ago
Sure, I implemented the paper "How to squeeze a lexicon" http://www.n3labs.com/pdf/lexicon-squeeze.pdf with a few tweaks a while back. It uses a simple fixed format for all nodes, which works very well for datasets which contains many com
49.
by gorset 15y ago
1 million pair using 16 MB is about 16 bytes per pair, which is perfectly fine but nothing impressive. The dataset is static, so a simple naive solution would be to create a big array sorted by key. Assuming both photo and user IDs use 4 by
50.
by gorset 15y ago
There's usually a tradeoff between many small pauses and few big ones (but even with CMS there will be a point where the JVM needs to do a full GC). The usual approach for low latency systems are to spend the GC pauses incrementally over ti
51.
by gorset 15y ago
The JVM is very good if you are able to handle or avoid the GC pauses. Unless you are running Azul's Zing JVM, some tuning is usually required for non-trivial systems. We have tuned our system to be very fast with low latency most of the ti
52.
by gorset 15y ago
That's actually accept_filter(9). The man page for freebsd has a interesting info: The accept filter concept was pioneered by David Filo at Yahoo! and refined to be a loadable module system by Alfred Perlstein. The closest you
53.
by gorset 15y ago
Yeah, the lack of support for TTYs can be annoying when writing a terminal application (a workaround for some cases is to use pipes), but it hardly qualifies as a significant problem for writing network applications. Here's a more interesti
54.
by gorset 15y ago
I fully agree that kqueue is awesome, but what specifically is broken on OSX? I've used in extensively on that platform, and haven't run into any showstoppers.
55.
by gorset 15y ago
Implementing approximate matching has been on my todo list for some time. I'm planning to either implement http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=5362... or http://www.sciencedirect.com/science/article/pii/S1570866704...
56.
by gorset 15y ago
If you don't need the add and remove operations, it's possible to reduce the size significantly by applying automata theory: http://www.n3labs.com/pdf/lexicon-squeeze.pdf It works extremely well "out of the box" for dictionary words, sinc
57.
by gorset 15y ago
Yes, exactly. I would love to have cross device SSO and authentication. Having proper cryptography available in the browser (either built-in or through a extension) would make that easier to make. As a bonus: phishing resistant.
58.
by gorset 15y ago
If you are running javascript on a page from a server, you are trusting the server with your plaintext.
59.
by gorset 15y ago
I want to use them for speeding up some critical code for compression/decompresison, the current code is not straightforward. How to implement the instructions in software is covered in Hacker's delight, and has some additions in the latest
60.
by gorset 15y ago
Wow, I can't wait to play around with PDEP and PEXT instructions.

More ›