This is hardly news but surely fun. Here is a link to Computer Languages Benchmark Game, the idea is to compare performance of different programming languages relative to each other. What follows are some of my casual, biased, ill-informed, overgeneralized observations (that's enough warning I guess)..
The Intel C++ compiler outperforms GNU C/C++ and even Intel C compiler. Is it just an experimental glitch or is C++ more amenable to optimization?
Comparing just the execution speed, Python Psyco (score 9.22) leaves standard Python biting dust (score 26.66). So JIT seems to be a big win. The results are impressive but not so much if you factor in the memory usage (scores 8.56 and 13.79 respectively). Similar pattern emerges for other JIT languages like Lua, but the differences are less pronounced.
Pascal comes very near to C/C++ in terms of performance, but nobody seems to use it these days (not even in teaching). Similar is the case with D (and this considering that some very smart people seem to be pushing it).
JVM languages except Java (Spidermonkey Javascript, JRuby) tend to cluster towards the bottom of the scorecard. Maybe their ultimate aim is not speed. Again, the differences become more pronounced if you neglect memory usage and just concentrate on speed. But maybe it has more to do with Static vs Dynamic languages than with JVM as such.
For nearly half of the languages mentioned, I can confidently claim that not a single line of their code has ever materialized before me. Reminds me of the frog in the pond.
So what is the lesson here? I don't know (But the game is fun).
The Intel C++ compiler outperforms GNU C/C++ and even Intel C compiler. Is it just an experimental glitch or is C++ more amenable to optimization?
Comparing just the execution speed, Python Psyco (score 9.22) leaves standard Python biting dust (score 26.66). So JIT seems to be a big win. The results are impressive but not so much if you factor in the memory usage (scores 8.56 and 13.79 respectively). Similar pattern emerges for other JIT languages like Lua, but the differences are less pronounced.
Pascal comes very near to C/C++ in terms of performance, but nobody seems to use it these days (not even in teaching). Similar is the case with D (and this considering that some very smart people seem to be pushing it).
JVM languages except Java (Spidermonkey Javascript, JRuby) tend to cluster towards the bottom of the scorecard. Maybe their ultimate aim is not speed. Again, the differences become more pronounced if you neglect memory usage and just concentrate on speed. But maybe it has more to do with Static vs Dynamic languages than with JVM as such.
For nearly half of the languages mentioned, I can confidently claim that not a single line of their code has ever materialized before me. Reminds me of the frog in the pond.
So what is the lesson here? I don't know (But the game is fun).
No comments:
Post a Comment