Speed comparisson between Kaffe/jdk on Linux/Windows

Constantin Teodorescu teo at flex.ro
Mon Apr 12 02:34:12 PDT 1999

Godmar Back wrote:
> Now, about Constantin's little test.  I assume he is talking about
> the iute.java program @http://rufus.w3.org/tools/Kaffe/messages/3796.html
> This program is no indicator of the quality of the jit compiler whatsoever.

Godmar is right.

I simply wrote that small program because I was curious how jdk 1.1.7 on
Windows is running with Symantec's JIT.
And I must say it was extremely fast. The Windows version has completed
the test in 2 seconds, the Linux one in 10 seconds.

> This is not to say that the results are uninfluenced by how good or bad
> the jit compiler is, but it is not a JIT benchmark.  What it really measures
> is the combination of the speed of the string and hashtable implementations
> and the allocator and garbage collector.

Yes. I quickly tried to emulate there a "real" program but I used only a
reduced set of operation (string operation, hashtables, math operations,
loops). It was NOT intented to be a benchmark. But it can be easily
extended and it may give an intuitive overall result of how fast a VM is
compared to another running the same program.

> Next, keep in mind that all performance results in Java are heavily
> influenced by the frequency and duration of garbage collection.
> Writing pure JIT benchmarks is hard.

I agree.

> To my knowledge, however, nobody has focused on writing such benchmarks;
> and the good reason for this is that they would be of little use

I cannot say that. I am writting for example a simple speed test program
just to decide what JVM I will use : jdk or kaffe.
Of course presuming that only the speed will be the decision.

I was watching kaffe for a long time. I am interested in kaffe because I
plan to use for an application as clients some "poor" machines
(486DX2/66 with only 4 Mb RAM) and I was excited reading about kaffe's
capabilities running java applications under pure DOS. The 1.0.b3 kaffe
was not able to run SWING code so I had to work with jdk 1.1.7 for

I am designing my applications with a free and simple patch to visual
Tcl that I wrote that allows visual Java development ( see
http://www.java.ro/vtclava/index.html ) . But my code is generating
SWING so I was more than pleased reading that the new released 1.0.b4 is
running also SWING code.

Best regards,
Constantin Teodorescu
FLEX Consulting Braila, ROMANIA

More information about the kaffe mailing list