Anyone who thinks C# or Java is going to execute faster than C++ needs their head read, but as pointed out above it will be the I/O operations which will be the actual bottlenecks.
Bingo. Also pointed out by DCraig. What makes me chuckle is that a lot of these gunslinger coders seem to have so little grasp of the architecture of the whole system. One wonders what the hell there code must be like. Guess all these guys have written there own lightweight email clients (C variant of course) so as to get there emails to there destinations quicker Cheers.
When your C/C++ app has core dumped with a segmentation violation, or been compromised by a buffer overflow exploit due to running off the end of an array, bad pointer arithmetic etc etc, and the Java app is still running having caught the exception and making do with what it has, then yes Java is faster - in fact infinitely faster. You might say that is poor coding on behalf of the C/C++ developer. Perhaps, but the world is full of programmers of ordinary ability and really good ones are quite rare. You might say also that tha C/C++ app was not sufficiently tested. But most testers are also of ordinary ability, test plans have holes in them and so on. Testing also costs money and time and there is a limit to testing. If the language protects against certain defects, or minimizes their seriousness, then that has to be a very good thing.
I'm not sure what purpose your little diatribe serves in relation to my post, the real problem is not the language, it the design of programs. In my experience the programmers who make the loudest noises about this language vs. that tend to be the ones who are actually the least capable designers. In the end you just have to use whatever language you are good at, just make sure your design is good. OO design principals apply to most modern languages.