Benchmark results for Linux JVM's.
Hello, We have formally evaluated the different virtual machines available for Linux and thought it would be worthwhile to share the results with the Linux community. The following tests were conducted on an unloaded dual processor Pentium II/400mhz running Debian GNU/Linux (kernel 2.2.8). Each benchmark execution was repeated ten times. We discarded the maximum and minimum results, and averaged the remaining 8 execution times. The first 9 benchmarks come from the specJVM98 benchmark suite (http://www.spec.org), and the last two benchmarks come from our own private collection. base(s): time in seconds to run under blackdown jdk 1.2, pre-release 2, with jit. sunint: speedup (base time/this time) of the blackdown jdk1.2, pre-release 2, with no jit. borjit: speedup of blackdown jdk1.2, pre-release 2, with the Borland jit installed (http://www.borland.com) ibmjit: speedup of the AlphaWorks IBM 1.1.8 JIT (http://www.alphaworks.ibm.com) A # indicates that the run failed validity checks. base(s)sunintborjitibmjit check .84 -1.33 -1.25 #1.75 - compress 65.61 - .15 -1.07 -2.42 - db 148.43 - .57 - .98 -2.98 - jack 64.50 - .43 -1.35 -3.65 - javac 75.67 - .54 -1.21 -2.51 - jess 50.86 - .47 -1.44 -2.67 - mpegaudio 54.61 - .15 -1.19 #2.32 - mtrt 40.32 - .41 -1.78 -2.79 - raytrace 55.56 - .45 -1.92 -3.04 - sablecc-w 42.57 - .58 -1.06 -2.32 - soot-j 132.93 - .69 -1.25 -2.26 - The conclusions are fairly obvious. Now, if only IBM had a jit for 1.2... We also evaluated shujit and tyajit, but they were unable to run most of the benchmarks correctly. Stay tuned for a comparison of NT Java Virtual Machines in the near future, on the same hardware. Permission is granted to re-distribute this e-mail in any medium as long as it remains unchanged. All trademarks belong to their respective owners. To everyone working on Java for Linux: keep up the great work! :) Best regards, Raja Vallee-Rai ([EMAIL PROTECTED]) Sable Research Group -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Benchmark results for Linux JVMs (formatted for 70 columns)
I goofed with the last message and messed up the formatting. Here's a better version: -Raja - Hello, We have formally evaluated the different virtual machines available for Linux and thought it would be worthwhile to share the results with the Linux community. The following tests were conducted on an unloaded dual processor Pentium II/400mhz running Debian GNU/Linux (kernel 2.2.8). Each benchmark execution was repeated ten times. We discarded the maximum and minimum results, and averaged the remaining 8 execution times. The first 9 benchmarks come from the specJVM98 benchmark suite (http://www.spec.org), and the last two benchmarks come from our own private collection. base(s): time in seconds to run under blackdown jdk 1.2, pre-release 2, with jit. sunint: speedup (base time/this time) of the blackdown jdk1.2, pre-release 2, with no jit. borjit: speedup of blackdown jdk1.2, pre-release 2, with the Borland jit installed (http://www.borland.com) ibmjit: speedup of the AlphaWorks IBM 1.1.8 JIT (http://www.alphaworks.ibm.com) A # indicates that the run failed validity checks. base(s)sunintborjitibmjit check .84 -1.33 -1.25 #1.75 - compress 65.61 - .15 -1.07 -2.42 - db 148.43 - .57 - .98 -2.98 - jack 64.50 - .43 -1.35 -3.65 - javac 75.67 - .54 -1.21 -2.51 - jess 50.86 - .47 -1.44 -2.67 - mpegaudio 54.61 - .15 -1.19 #2.32 - mtrt 40.32 - .41 -1.78 -2.79 - raytrace 55.56 - .45 -1.92 -3.04 - sablecc-w 42.57 - .58 -1.06 -2.32 - soot-j 132.93 - .69 -1.25 -2.26 - The conclusions are fairly obvious. Now, if only IBM had a jit for 1.2... We also evaluated shujit and tyajit, but they were unable to run most of the benchmarks correctly. Stay tuned for a comparison of NT Java Virtual Machines in the near future, on the same hardware. Permission is granted to re-distribute this e-mail in any medium as long as it remains unchanged. All trademarks belong to their respective owners. To everyone working on Java for Linux: keep up the great work! :) Best regards, Raja Vallee-Rai ([EMAIL PROTECTED]) Sable Research Group -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Clarifications on the benchmark results.
Hello again, I took for granted the notation that I used in the results. It might be confusing for anyone who hasn't seen it before. Here are some explanations: The first column are in seconds, and are the execution times for the benchmarks under the Sun JIT. All the other numbers are ratios comparing their execution time to this base execution times. This ratio (base time / new time ) can be called "speedup". So a speedup of 2.00 means that it's twice as fast. For example: compress takes 65.61s to run with the Sun JIT. With the Sun intepreter its speedup is .15, so that's a really big slow down (more than 6 times slower). But with IBM's jit the speedup is 2.42. So IBM's JIT is more than twice as fast as Sun'sJIT. This should clear up any confusion. Raja -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
