Mono performance

amazedsaint > There is an interesting post from Miguel on this.

That chart is way out of date!

Here's the current Mono C# 4.0 comparison

There is an interesting post from Miguel on this.

"Alvaro points out that in the Language Shootout Benchmark Mono comes in 18th place compared to Java's 10th place.

We know that Sun's proprietary Java edition (not the open source one, as that one is nowhere to be found yet) is faster than Mono, but I was surprised that we were so far behind. So I looked at the comparison between Java6 and Mono."

_{(source: tirania.org)}

See Details Here

I benchmarked Mono 2.2 against various other VMs including OCaml, LLVM, .NET and Java:

Mono 2.2 vs OCaml vs .NET vs LLVM vs JDK http://2.bp.blogspot.com/_NMRkpon4Ps0/SYNY8YtzTCI/AAAAAAAAACg/Bkjli4JV-NU/s1600-h/scimark2_individual.gif

In summary, Mono 2.2 is better than Mono 2.0 but is still usually a lot slower than almost everything else.

The latest version of this shoot-out can be found here.

In a nutshell, compared to Java, Mono uses much less memory and runs a little slower.

To keep things in perspective, they both run about the same speed in the grand scheme of things. They both run these tests orders of magnitude faster than languages like Ruby, LUA, Python, PHP, and Perl. They are both in the same ball-park as Pascal, OCaml, Haskell, and LISP which makes them about half as fast as C and C++. Click here for more detail.

To nitpick, we should be clear that our answers here compare Java and Mono C#. Both Mono and the JVM are environments targeted by various languages and the choice of language matters. For example, if we compare F# on Mono with Clojure on the JVM we get totally different results and suddenly "Mono" is "faster".

While these comparisons are interesting and informative, you have to take them with a grain of salt. First, although some people do, most people writing Java and C# code are not running nothing but numerical algorithms in tight loops. Second, in some of these languages, the tested algorithms may just be calling into libraries that are written in C and not in the target language at all. Finally, other benchmarks (and implementations) would likely yield different results.

Still, it is reasonable to infer that Go is much faster than Ruby, C++ is faster than Go, and both Java and C# perform pretty reasonably given how productive they are (IMHO).