Performance Improvements Using Judy Arrays
1–10 of 53 posts
Re: Performance Improvements Using Judy Arrays
#2This is great. Someone should change the title so it's descriptive of what's going on, although I have no idea what that title should be, heh.
This post is about Github moving some code from ruby to C for performance, MRI's GC, and judy trie-like data structure and how it compares wrt to both performance and memory consuption with hash implementations.
Re: Performance Improvements Using Judy Arrays
#3Some previous discussion on Judy Arrays [0]. Apparently they are patented!
Re: Performance Improvements Using Judy Arrays
#4This is great. Someone should change the title so it's descriptive of what's going on, although I have no idea what that title should be, heh. This post is about Github moving some code from ruby to C for performance, MRI's GC, and judy trie-like data structure and how it compares wrt to both performance and memory consuption with hash implementations.
"... I have no idea what that title should be, heh."
It's a reference to the Beatles' song, "Hey Jude" (which opens with "Hey, Jude, don't make it bad | Take a sad song and make it better ...").[1]
[1] http://en.wikipedia.org/wiki/Hey_Jude
Edited for clarity.
Re: Performance Improvements Using Judy Arrays
#5Wow this is an incredibly interesting and accessible article on performance tuning. I have two questions:
1) Is there an easy tutorial somewhere on calling out to native code from Ruby?
2) Could the team at Github possible give a little more detail on what you did to get all of those pretty benchmarking graphs? How did you get all of the info on memory usage and CPU activity (I assume it wasn't just time {command} > file.txt)
Re: Performance Improvements Using Judy Arrays
#6Something here doesn't sit right with me. Please don't launch into a lecture on algorithmic complexity and cache lines if you are running on Ruby on Rails.
Re: Performance Improvements Using Judy Arrays
#7This is great. Someone should change the title so it's descriptive of what's going on, although I have no idea what that title should be, heh. This post is about Github moving some code from ruby to C for performance, MRI's GC, and judy trie-like data structure and how it compares wrt to both performance and memory consuption with hash implementations.
"... I have no idea what that title should be, heh." It's a reference to the Beatles' song, "Hey Jude" (which opens with "Hey, Jude, don't make it bad | Take a sad song and make it better ...").[1] [1] http://en.wikipedia.org/wiki/Hey_Jude Edited for clarity.
The issue is that the title, while clever, doesn't give a even a remote idea of what the post is about.
Re: Performance Improvements Using Judy Arrays
#8Earlier quoted context omitted.
"... I have no idea what that title should be, heh." It's a reference to the Beatles' song, "Hey Jude" (which opens with "Hey, Jude, don't make it bad | Take a sad song and make it better ...").[1] [1] http://en.wikipedia.org/wiki/Hey_Jude Edited for clarity.
The issue is that the title, while clever, doesn't give a even a remote idea of what the post is about.
I guess I still prefer a catchy title to preachy ones that extensively use RFC 2119 keywords. e.g. "Why you SHOULD profile Ruby code to catch expensive regex bottlenecks" :)
Personal preference though.
Re: Performance Improvements Using Judy Arrays
#9I would have liked to seen the Judy array implementation in pure Ruby, so we could compare apples to apples. I'm not trying to troll... but basically they solved their problem by:
* Avoiding Ruby language features
* Rewriting it in a different language
This is why I lean towards static languages like Go, Scala, and Java.
Re: Performance Improvements Using Judy Arrays
#10Judy arrays are also faster in node.js: http://legitimatesounding.com/blog/Faster_sometimes_Associat...
the issue is with the patents - has anyone been contacted yet about the usage?