Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
zhenjl
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
11 ms
·
61.
▲
Graceful Shutdown of Go net.Listeners
(zhen.org)
5 points
by
zhenjl
13y ago
|
0 comments
62.
▲
Ring Buffer - Variable-Length, Low-Latency, Lock-Free, Disruptor-Style
(zhen.org)
3 points
by
zhenjl
13y ago
|
0 comments
63.
▲
by
zhenjl
13y ago
I did not. It might be difficult tho for all the delta encoding methods since they all depend on computing delta from the previous integer. For the non-delta methods it is conceivable that ||ism may help, but not sure by how much.
64.
▲
by
zhenjl
13y ago
2013-11-17 Update #2: Tried another new trick. This time I updated the leading bit position function, when using the gccgo compiler, to use libgcc’s __clzdi2 routine. This had the same effect as the update #1 except it’s for when gccgo is u
65.
▲
by
zhenjl
13y ago
Ok, I think I definitely fixed it now. Thanks for the heads up.
66.
▲
by
zhenjl
13y ago
I just finished updating the code with an assembly version of a hot function (finding the leading "1" bit position) and updated the blog to reflect the improvements. In general I saw a 3-40% performance increase using the assembly
67.
▲
by
zhenjl
13y ago
bsg75, thanks for this. I updated to serve images from the site instead of from Google. Hopefully this helps. Please let me know if you still run into problems. I haven't updated the data section yet.
68.
▲
by
zhenjl
13y ago
Interesting. Those are served from Google. Not sure what's going on there. Anyone else have this problem?
69.
▲
Go vs Java: Decoding Billions of Integers Per Second
(zhen.org)
21 points
by
zhenjl
13y ago
|
13 comments
70.
▲
by
zhenjl
13y ago
Thanks for the suggestion. I resubmitted here: https://news.ycombinator.com/item?id=6710115
71.
▲
(Resubmit) Improving Cityhash Performance by Go Profiling
(zhen.org)
6 points
by
zhenjl
13y ago
|
0 comments
72.
▲
by
zhenjl
13y ago
Thanks for the heads up. Wow...can't believe I missed that. Will have to resubmit.
73.
▲
Improving Cithhash Performance by Go Profiling
(zhen.org)
2 points
by
zhenjl
13y ago
|
3 comments
74.
▲
by
zhenjl
13y ago
Sorry for the late reply. I was traveling for the past week or so. In any case, EWAH is designed to avoid the WAH patent. According to the author, the Apache people checked things out when they adopted it for Apache Hive and they were appa
75.
▲
by
zhenjl
13y ago
The algorithm is EWAH, not WAH. It's also by Daniel lemire.
76.
▲
by
zhenjl
13y ago
Thanks for commenting! And you are absolutely correct about the bar charts. Not exactly sure why I chose line but they have been replaced now. The testing didn't use the built-in benchmark mechanisms but it did use the *testing.T mecha
77.
▲
Benchmarking Integer Compression in Go
(zhen.org)
5 points
by
zhenjl
13y ago
|
6 comments
78.
▲
Bitmap Compression using EWAH in Go
(zhen.org)
1 points
by
zhenjl
13y ago
|
0 comments
79.
▲
by
zhenjl
13y ago
Author here. (same comment posted to reddit) @bradfitz thanks for the comments and suggestion. I updated the code based on your suggestion (not allocate for every add/check) and the performance improved by 25+%!! Appreciate it! I updat
80.
▲
by
zhenjl
13y ago
Sorry for the delay in responding there. Hadn't check this for a couple of days. You are correct. In fact, I am allocating in both standard and partitioned Check() implementations, in the bits() func. I can pre-allocate during New() an
81.
▲
by
zhenjl
13y ago
Austin, yeah pretty sure it's broken, unfortunately. I used an implementation another person ported but results seem to come out weird. I need to figure out why when I get a moment, or do a port of my own.
82.
▲
Benchmarking Bloom Filters and Hash Functions in Go
(zhen.org)
3 points
by
zhenjl
13y ago
|
5 comments
83.
▲
Go Skiplist - alternative to binary search trees
(zhen.org)
4 points
by
zhenjl
13y ago
|
0 comments
84.
▲
by
zhenjl
13y ago
Good points. Both thrift and protobuf (and Avro) seemed heavy weight due to their requirements. We wanted something lightweight, thus BSON and MsgPack. In terms of rewriting a bunch of code, msgpack and bson both easily marshall/unmars
85.
▲
Testing MsgPack and BSON
(zhen.org)
4 points
by
zhenjl
13y ago
|
3 comments
86.
▲
Migrating 40,000 Posts from Wordpress to Octopress
(zhen.org)
1 points
by
zhenjl
13y ago
|
0 comments
87.
▲
by
zhenjl
13y ago
just ported cityhash's test program over to Go and passed!! Type casting like this makes my brain hurt. b = uint32(int64(b) * int64(c1) + int64(int8(v)))
88.
▲
Show HN: Google City Hash re-implemented in Go
(github.com)
3 points
by
zhenjl
13y ago
|
1 comments
89.
▲
by
zhenjl
13y ago
Hm..is it me or there's no way to switch branch using the mobile view..
90.
▲
Who are the biggest and baddest cloud providers?
(topclouds.org)
1 points
by
zhenjl
13y ago
|
0 comments
More ›