Hi,
I can confirm Jerome's results. I tried out the test with 3
different optimization options (Optimize for size, for speed, and
'maximum optimizations') for the MSVC 7.1 (2003) compiler. In all the
cases, the running time was 0 seconds. When I compiled and ran the
code with no optimizations, I got a running time of 3.1090 seconds.
For a more comprehensive test of ICC, GCC and MSCV, take a look at
the following page:
- Performance Comparison of Java/.NET Runtimes (Oct 2004)
(http://www.shudo.net/jit/perf/)
Found the link on the Harmony wiki, :) under the JVM Benchmarks
section (http://wiki.apache.org/harmony/JVM_Benchmarks).
The test pits GCC 3.4.2 against ICC 8.1 and MSVC 7.1. So maybe we
should re-run the tests in two months, when we have MSVC 8 to test
against GCC 4.0.2 and ICC 9. ;)
Regards,
Tanuj
On 10/17/05, Enrico Migliore <[EMAIL PROTECTED]> wrote:
> Jeroen Frijters wrote:
>
> >Enrico Migliore wrote:
> >
> >
> >> the code is a simple function that gets called 300000000 times.
> >>
> >>
> >
> >This is a flawed test. Your code doesn't actually do anything, so the
> >function can be optimized away entirely and this is in fact was MSVC 7.1
> >does ("the test lasted for 0.0000 seconds")
> >
> >Regards,
> >Jeroen
> >
> >
> >
> >
> >
> Hi Jeroen
>
> to make an exhaustive comparison of the three tool chains I certainly
> need a test suite.
>
> In this simple test, I just wanted to see how the compilers build up the
> stack frame of function calls,
> and consequently how fast is the context change.
>
> Enrico
>
>
