fannkuch-redux benchmark N=12

Each chart bar shows how many times slower, one ↓ fannkuch-redux program was, compared to the fastest program.

These are not the only programs that could be written. These are not the only compilers and interpreters. These are not the only programming languages.

Column × shows how many times more each program used compared to the benchmark program that used least.

    sort sortsort
  ×   Program Source Code CPU secs Elapsed secs Memory KB Code B ≈ CPU Load
1.0Ada 2005 GNAT #3 38.259.611,9762100  100% 99% 100% 100%
1.3Fortran Intel #3 51.2012.848,5641148  100% 100% 100% 100%
1.4C++ g++ #5 53.3313.631,2641440  100% 98% 99% 94%
1.4C++ g++ #4 53.5613.731,2641439  99% 100% 98% 94%
1.5C gcc #2 55.4814.254281557  98% 96% 96% 100%
1.7Haskell GHC #5 63.3916.154,036834  99% 100% 100% 94%
1.7OCaml #4 0.0016.3011,2801004  100% 100% 100% 100%
1.7Lisp SBCL #4 65.1216.7534,0041518  98% 99% 100% 92%
1.8Go 67.1616.861,032900  100% 100% 100% 100%
1.8OCaml #3 0.0117.3621,3721017  100% 100% 100% 100%
1.8Java  70.0617.7319,7201282  99% 99% 98% 99%
1.8Scala #2 69.8917.7725,6401017  97% 98% 100% 98%
1.9Rust 71.7817.9939,844644  100% 100% 100% 100%
2.2Clojure #3 82.0121.3359,1761491  99% 95% 96% 96%
2.3Pascal Free Pascal 87.2821.877521018  100% 100% 100% 100%
2.3C# Mono #3 87.7322.3122,9841096  99% 99% 98% 98%
2.3F# Mono #3 84.1622.4130,140945  95% 94% 94% 93%
2.4C gcc #4 22.6622.673921183  0% 1% 100% 0%
2.4C++ g++ #7 22.6922.703601150  1% 0% 1% 100%
5.1C# Mono #2 48.9949.0121,608564  1% 0% 1% 100%
5.4Clojure #2 166.4852.0762,9601088  80% 80% 79% 79%
5.5C gcc #3 52.4452.47392567  1% 99% 1% 2%
6.1Fortran Intel 58.8258.84520590  0% 0% 0% 100%
6.2C++ g++ #3 59.1559.17360593  1% 100% 0% 0%
6.2Lisp SBCL #5 59.4659.4834,456674  1% 0% 1% 100%
6.6C gcc 63.1763.19392508  1% 1% 100% 0%
6.6F# Mono #4 63.7563.7722,376612  100% 1% 0% 1%
6.9Haskell GHC #4 71.1466.002,460658  3% 3% 3% 100%
7.2Java  #2 69.6169.5719,664514  0% 99% 1% 1%
7.3C# Mono 69.9169.9419,308520  1% 100% 0% 0%
7.6Racket #3 286.8072.7320,6761096  100% 95% 100% 100%
7.7Lisp SBCL #3 74.1474.1723,504821  100% 0% 0% 1%
8.3Dart #2 80.6679.85171,376495  1% 1% 1% 100%
8.8F# Mono #2 84.3884.4120,940548  1% 100% 1% 0%
9.2Scala 88.7688.7023,636459  1% 1% 99% 0%
9.4OCaml #2 90.3990.42764473  93% 0% 0% 7%
11OCaml 106.38106.41764524  96% 0% 0% 4%
13Erlang HiPE 8 min125.7312,8721038  99% 100% 97% 97%
15F# Mono 144.31144.1824,612551  39% 23% 25% 15%
17Haskell GHC #2 8 min162.275,344808  83% 83% 83% 82%
20Lisp SBCL #2 188.07188.3878,600513  49% 1% 1% 51%
21Erlang 13 min203.7611,7961038  100% 98% 100% 100%
28Racket #2 272.16272.1420,220903  0% 1% 100% 0%
31Racket 297.55297.5218,840649  1% 41% 0% 59%
51Ruby JRuby #2 30 min8 min650,4721426  98% 91% 94% 90%
55Perl #2 35 min8 min8,720565  100% 99% 99% 99%
56Haskell GHC 12 min8 min3,948553  55% 28% 32% 25%
80Ruby JRuby 12 min12 min632,076384  33% 21% 24% 24%
84Ruby #2 52 min13 min31,5121426  98% 97% 97% 98%
90PHP #3 56 min14 min14,9681150  99% 98% 99% 95%
95Python 3 #2 59 min15 min40,096797  98% 97% 100% 97%
104Python 3 1h 05 min16 min40,1001108  99% 100% 99% 100%
104Hack 16 min16 min55,344532  1% 0% 0% 100%
243Python 3 #6 38 min38 min5,428385  1% 0% 0% 100%
253Ruby 40 min40 min7,508384  0% 0% 0% 100%
272Perl 43 min43 min1,864457  0% 0% 0% 100%
305PHP #2 48 min48 min3,256441  100% 0% 0% 0%
356PHP 57 min57 min3,312482  3% 97% 0% 0%
C++ g++ Failed1059
Haskell GHC #3 Timed Out5 min1153
Hack #2 Timed Out1h 20 min440
Hack #3 Failed1150
"wrong" (different) algorithm / less comparable programs
1.1Java  #3 40.9810.4121,0321633
1.2C# Mono #5 46.9111.9924,7801400
1.4Lisp SBCL 52.3813.2843,6841607
3.9C++ g++ #6 37.8637.87364894
4.2C# Mono #4 40.2640.2719,572710

 fannkuch-redux benchmark : Indexed-access to tiny integer-sequence

diff program output N = 7 with this output file to check your program is correct before contributing.

We are trying to show the performance of various programming language implementations - so we ask that contributed programs not only give the correct result, but also use the same algorithm to calculate that result.

For N = 7 programs should generate these permutations (40KB) - which, incidentally, seem to be in the same order as permutations generated by the Tompkins-Paige algorithm, see pages 150-151 Permutation Generation Methods Robert Sedgewick.

The fannkuch benchmark is defined by programs in Performing Lisp Analysis of the FANNKUCH Benchmark, Kenneth R. Anderson and Duane Rettig.

Each program should

The conjecture is that this maximum count is approximated by n*log(n) when n goes to infinity.

FANNKUCH is an abbreviation for the German word Pfannkuchen, or pancakes, in analogy to flipping pancakes.


Thanks to Oleg Mazurov for insisting on a checksum and providing this helpful description of the approach he took -

Revised BSD license

  Home   Conclusions   License   Play