Difference between revisions of "2008 LightningTalks"
From JVMLangSummit
Jump to navigationJump to searchLine 24: | Line 24: | ||
|Day Two, 4:45 | |Day Two, 4:45 | ||
|Why is your code slow? | |Why is your code slow? | ||
− | |Debugged lack-of-inlining in JRuby and lack-of-loop-unrolling in ASM in about 1/2hr each. Fixnum/BigNum costs are high, but perhaps not for why you might expect (it's NOT the GC costs, it's allocation and final fields). | + | |Debugged lack-of-inlining in JRuby and lack-of-loop-unrolling in ASM in about 1/2hr each. Fixnum/BigNum costs are high, but perhaps not for why you might expect (it's NOT the GC costs, it's allocation and final fields). Academic accounts available upon request; email cliffc@azulsystems.com |
|} | |} |
Revision as of 14:49, 25 September 2008
Lightning talks are held in the afternoon (~16:30-17:30).
Title | Presenter | Time & Place | Audience | Summary |
---|---|---|---|---|
Optimizing for Agility | Stuart Halloway | Day One, 4:30? | People who prioritize developer performance over language performance | Agility requires smooth adaptation to change. With multiple JVM languages to choose from, you should pick an agile language for most dev, and drop to a performant language occasionally. I'll tell you which is which. :-) |
Language Addicts Anonymous | John D. Mitchell | Day Two, 4:45 | Humans and meat-puppets | As language and platform creators, are we actually attacking and solving the real problems? Or, are we addicted to playing in the same old ghetto? |
Azul Performance Analysis Tool | Cliff Click | Day Two, 4:45 | Why is your code slow? | Debugged lack-of-inlining in JRuby and lack-of-loop-unrolling in ASM in about 1/2hr each. Fixnum/BigNum costs are high, but perhaps not for why you might expect (it's NOT the GC costs, it's allocation and final fields). Academic accounts available upon request; email cliffc@azulsystems.com |