They aren’t the ones bashing anyone, you are bashing them.

This time you’ve chosen the ’speed/performance’-comparison as a tool, and you have failed miserably.

Note: comment has been flagged for arriving from a possible incarnation of a known (eet), pseudonymous, forever-nymshifting, abusive Internet troll that posts from open proxies and relays around the world.

(actually, I’m just bumping to further humiliate Roy, but I can pretend that I’m actually interested, can’t I?)

Nice try at changing the subject, Roy, but it was an epic fail.

Even if Mono was the antichrist, it still wouldn’t change the results of the tests.

http://www.microsoft.com/interop/msnovellcollab/moonlight.mspx

Care to explain what this means if one didn’t pay Novell (Microsoft’s software patent royalties)?

Read here for: an failed attempt at reproducing his results, courtesy of Jeffrey Steadfast.: http://jeffreystedfast.blogspot.com/2008/05/debian-language-benchmarks-sumfile.html

Before accusing Mr. Steadfast of an outright lie (which I espect of you anti-Mono extremists): The data is there, the source is there; compile it and reproduce before you make any such claims!

1. that the brain-dead ReadLine() + Int32.Parse() implementation of THIS PARTICULAR TEST in C#/Mono can run as fast as the same brain-dead ReadLine() + Integer.Parse() implementation in Java. (as in, comparing apples to apples)

2. that he can write a much more optimized C# implementation that can outperform the fgets_unlocked() + atoi() implementation in C (which he successfully proved he could).

You also misrepresent the facts when you claim that he was bashing Java. Nowhere in his blog post did he bash Java. Nowhere.

You are scum for suggesting otherwise.

You refused to run the tests yourself and so AlexH offered, to which you claimed he was biased and so results from his reproduction couldn’t be trusted either.

Are you now saying that you trust that the results are indeed accurate and that the guy’s C# implementation was 6-7x faster than the fastest C implementation on the Debian Language Shootout site?

You are just trying to find excuses now, pretty pathetic.

All these things are implementation/program-dependent. One could prove almost everything that’s desirable, so benchmarks are a more complex things than that. I’m a technical researcher and I know that peer review would scrutinize for exactly this reason. One quote that also comes to mind:

“Microsoft did sponsor the benchmark testing and the NT server was better tuned than the Linux one.”

http://www.itweb.co.za/sections/enterprise/1999/9904221410.asp

I’ll be waiting.

You fell victim to one of the classic blunders! The most famous is never get involved in a land war in Asia, but only slightly less well-known is this: never go in against fejj when parsers or I/O performance is on the line! Ha ha ha ha ha ha ha! Ha ha ha ha ha ha ha! Ha ha ha…

I might have to add his blog to my rss feed.

[miles@localhost ~]$ time mono sumcol2.exe < sumcol-input100000.txt
50000000

real 0m3.799s
user 0m3.476s
sys 0m0.316s
[miles@localhost ~]$ time ./sumcol < sumcol-input100000.txt
50000000

real 0m18.303s
user 0m18.001s
sys 0m0.284s

(sorry for the repost, had to html encode the less-than char)

[miles@localhost ~]$ time mono sumcol2.exe