Java speeds have caught up to C/C++ and sometimes passes them

[MK27]

I think what's possibly more relevant WRT to the difference between comparable applications done in C, C++, and Java is the amount of memory consumed. The benchmark game sort of hides from this because the tasks are simple one-off things, but with large and complex programs you'll almost certainly notice the one done in C uses substantially less memory than the one done in C++ or Java, if the C++ or Java programmer is really taking advantage of all the features of the language. Of course, the code base for the C version will probably be bigger too (due to the lack of features ).

[igouy]

The benchmark game sort of hides from this because the tasks are simple one-off things, but with large and complex programs you'll almost certainly notice the one done in C uses substantially less memory than the one done in C++ or Java, if the C++ or Java programmer is really taking advantage of all the features of the language. Of course, the code base for the C version will probably be bigger too (due to the lack of features ).

With respect to Java, the benchmarks game exaggerates Java memory use "because the tasks are simple one-off things" and unlike those tasks "with large and complex programs you'll almost certainly" need to allocate more memory than the JVM default ;-)

So instead of the n-body task where the Java program is listed as using 47x more memory than the C program, for memory usage we should look at k-nucleotide, reverse-complement, binary-trees and regex-dna - Java programs using 2-3x more memory than the corresponding C programs.