x64 Ubuntu : Intel® Q6600® quad-core
Computer Language Benchmarks Game      Help

- all benchmarks - binary-trees chameneos-redux fannkuch fasta k-nucleotide mandelbrot meteor-contest n-body pidigits regex-dna reverse-complement spectral-norm thread-ring - all languages - Ada 2005 GNAT ATS C GNU gcc C# Mono C++ GNU g++ Clean Erlang HiPE F# Mono Fortran Intel Go 6g 8g Haskell GHC Java 6 -server Java 6 -Xint Java 6 steady state JavaScript TraceMonkey JavaScript V8 Lisp SBCL Lua OCaml Pascal Free Pascal Perl PHP Python CPython Ruby JRuby Ruby MRI Scala Scheme PLT Smalltalk VisualWorks

"For every complex problem there is an answer that is clear, simple, and wrong."  H. L. Mencken

What fun! Increase the Code multiplier! Ignore Memory! Zero out those benchmarks that push your favourite language to the bottom!

Which balance between Code-used and Time-used and Memory-used is best for you? Zero out those benchmarks which are unlike what your programs do!

×   language implementation GM missing 1.0 C++ GNU g++ 1.22   1.3 C GNU gcc 1.60   1.6 Ada 2005 GNAT 1.92 1 1.7 Java 6 steady state 2.02 4 1.7 ATS 2.10 1 1.8 Java 6 -server 2.24   2.2 Haskell GHC 2.68   2.4 Scala 2.91   3.0 Fortran Intel 3.61 5 3.1 Lisp SBCL 3.76 1 3.5 OCaml 4.30   3.5 C# Mono 4.31   4.7 Pascal Free Pascal 5.74 1 4.8 Clean 5.80 4 7.3 F# Mono 8.94 3 8.4 Go 6g 8g 10.25   11 Erlang HiPE 13.06   13 Java 6 -Xint 15.32   14 JavaScript V8 17.33 4 18 Scheme PLT 21.81   23 Lua 27.64 2 25 PHP 31.02 2 28 Perl 34.36 1 29 Python CPython 35.86   33 Smalltalk VisualWorks 40.18 3 60 Ruby JRuby 73.53   100 Ruby MRI 122.59

data for largest N full data - smaller and largest N multipliers Time secs Memory KB Code B benchmark weight binary-trees chameneos-redux fannkuch fasta k-nucleotide mandelbrot meteor-contest n-body pidigits regex-dna reverse-complement spectral-norm thread-ring

 Which is best for you?

Which balance between Code-used and Time-used and Memory-used is best for you?

Read "How not to lie with statistics: the correct way to summarize benchmark results" (pdf).

For each benchmark, B is the best measurement; for each language implementation, the best measurement L is then normalized to ^L/_B

GM is the Weighted Geometric Mean of those ^L/_B ratios.

missing : Language implementations with more than a couple of Timeouts (at the largest workload) distort the ranking - so they have been excluded. They can still be compared directly against another language implementation.

There are other ways to analyse and present this data: ask Which programming languages have the fastest programs? or look for patterns in Code-used Time-used Shapes or take the Summary Data and do your own analysis!