Thursday February 22, 2007 There is an article that debunks the rumor that Java is slow. This article suggests that the "myth" of C/C++ applications, as native binary, will always be faster than Java, is not true. To a certain extend I agree with the author - poorly written C/C++ app can perform much worse than a Java app. Now, Java is my favorite programming language, but as someone working in HPC, all I can say is that I'm not at all convinced.
Posted by melvin ( Feb 22 2007, 04:53:26 PM SGT ) Permalink Comments [4]
This is a personal weblog, I do not speak for my employer.
| « December 2009 | ||||||
| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
|---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 | 31 | ||
| Today | ||||||
Today's Page Hits: 43
Right now I am trying to tune a bit of performance-critical financial/numeric C++ code for a client. The (GCC3.2.x) static compiler has done a good job, but cannot take care of about ~25% non-numeric overhead testing and retesting the same conditions, using virtual calls though there is no actual run-time polymorphism, etc, that HotSpot would be able to detect and optimise away at run-time.
Rgds
Damon
Posted by Damon Hart-Davis on February 22, 2007 at 07:44 PM SGT #
Posted by Brian Utterback on February 22, 2007 at 10:58 PM SGT #
Posted by 203.125.39.114 on February 23, 2007 at 10:05 AM SGT #
Posted by 217.86.125.164 on April 01, 2007 at 12:41 AM SGT #