C++ vs. C# vs. Assembly Language

[Michael A. Covington]

Re: C++ vs. C# vs. Assembly Language


"Jon Skeet [C# MVP]" <skeet@pobox.co m> wrote in message
news:MPG.19f6d1 d6370e16a598988 8@msnews.micros oft.com...[color=blue]
> Jason Smith <jason@nospam.c om> wrote:[color=green]
> > Some numbers:
> >
> > I've been playing around with a program for solving the Rubik's Cube[/color][/color]
using[color=blue][color=green]
> > brute force. I wrote it in Java, C# using safe code, and C# using[/color][/color]
pointers,[color=blue][color=green]
> > each one faster than the last.
> >
> > C# using pointers: 6x
> > C# using safe code: 3x
> > Java (Sun 1.4) using code optimized for speed: 1x
> >
> > I haven't written it in C/C++, but seeing how much faster it is than Sun
> > Java for this particular task, I have to think it is getting in the same
> > ballpark.[/color]
>
> I'd be interested to see the code you're using there - C# and Java are
> roughly the same speed in tests I've done. C# certainly isn't 3 times
> as fast "in general".[/color]

How about Prolog? If you want fast brute-force searches...

https://www.ai.uga.edu/mc/courses.html

[Anthony Berglas]

Re: C++ vs. C# vs. Assembly Language

Did you use the sever optimized Java? In my tests it is substantially
faster than .Net. (The default client Java was slower than .Net, I wonder
why they still ship both.) Also in my test Java beat C++ because it did
some clever inlining.

It would be interesting to get some idea as to why you see the differences.

Anthony

"Jon Skeet [C# MVP]" <skeet@pobox.co m> wrote in message
news:MPG.19f6d1 d6370e16a598988 8@msnews.micros oft.com...[color=blue]
> Jason Smith <jason@nospam.c om> wrote:[color=green]
> > Some numbers:
> >
> > I've been playing around with a program for solving the Rubik's Cube[/color][/color]
using[color=blue][color=green]
> > brute force. I wrote it in Java, C# using safe code, and C# using[/color][/color]
pointers,[color=blue][color=green]
> > each one faster than the last.
> >
> > C# using pointers: 6x
> > C# using safe code: 3x
> > Java (Sun 1.4) using code optimized for speed: 1x
> >
> > I haven't written it in C/C++, but seeing how much faster it is than Sun
> > Java for this particular task, I have to think it is getting in the same
> > ballpark.[/color]
>
> I'd be interested to see the code you're using there - C# and Java are
> roughly the same speed in tests I've done. C# certainly isn't 3 times
> as fast "in general".
>
> --
> Jon Skeet - <skeet@pobox.co m>
> http://www.pobox.com/~skeet
> If replying to the group, please do not mail me too[/color]

[Jon Skeet [C# MVP]]

Re: C++ vs. C# vs. Assembly Language

<"Anthony Berglas" <aberglas at frontsys (a spam free .com)>> wrote:[color=blue]
> Did you use the sever optimized Java? In my tests it is substantially
> faster than .Net. (The default client Java was slower than .Net, I wonder
> why they still ship both.)[/color]

The client JVM starts up faster and is generally better for short-
running processes. The server JVM is generally better for long-running
processes.
[color=blue]
> Also in my test Java beat C++ because it did
> some clever inlining.
>
> It would be interesting to get some idea as to why you see the differences.[/color]

It really depends on exactly what you're doing. For some operations,
the .NET JIT will beat the HotSpot JIT hands down - and for some
operations, the reverse is true. They have some pretty radically
different assumptions, to start with - HotSpot will progressively JIT
with more and more optimisations as it sees one section of code being
used more and more, whereas the .NET JIT is a "one off" JIT. Sometimes
one way works better, sometimes the other does.

--
Jon Skeet - <skeet@pobox.co m>

http://www.pobox.com/~skeet

If replying to the group, please do not mail me too