How big is the measured performance difference?

Each chart bar shows how many times more Time or how many times more Memory one unidentified ↓ spectral-norm program used, compared to the program that used least Time or the program that used least Memory.

 spectral-norm benchmark N=5,500

This table shows 5 measurements - CPU Time, Elapsed Time, Memory, Code and ~ CPU Load.

Compare how much Memory the spectral-norm programs used - sort Memory KB. Compare how much Code the programs used - sort Code B

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

    sort sort sort sort
  ×   Program Source Code CPU secs Elapsed secs Memory KB Code B ~ CPU Load
1.0GNU gcc #4 11.872.997721139  99% 100% 99% 99%
1.0C++ GNU g++ #7 11.892.991,1961114  100% 99% 99% 99%
1.3Ada 2005 GNAT #3 15.693.982,5281702  98% 99% 98% 99%
1.3Fortran Intel 15.934.001,276568  99% 99% 100% 100%
1.4Haskell GHC 16.024.112,260869  96% 99% 96% 99%
1.4Java 6 steady state #2 17.144.3124,6201027  99% 99% 99% 99%
1.5Java 6 -server #2 17.334.5115,208950  98% 95% 94% 97%
1.5Scala #2 17.664.5620,720720  96% 96% 97% 98%
1.6Ada 2005 GNAT #2 18.754.743,0121464  99% 99% 99% 98%
1.9C# Mono #2 22.295.625,3681063  99% 99% 99% 99%
2.0ATS #2 22.065.961,6562339  92% 92% 93% 93%
2.0Lisp SBCL #3 22.226.017,476883  92% 92% 93% 93%
2.1OCaml #3 20.026.213,304907  79% 79% 81% 81%
2.2Go 6g 8g #2 26.346.591,472545  100% 100% 100% 100%
4.1Erlang HiPE #2 47.7012.1813,348747  98% 98% 97% 98%
5.3Fortran Intel #2 15.7015.701,100513  0% 0% 100% 0%
5.3Lisp SBCL 15.7315.728,524625  0% 0% 100% 0%
5.3Haskell GHC #3 15.9415.941,812737  0% 0% 0% 100%
5.5Java 6 -server 16.5716.4215,908514  100% 0% 1% 0%
5.5Scala 16.5716.4316,560500  0% 0% 100% 0%
6.3Pascal Free Pascal 18.7918.79192423  0% 0% 0% 100%
6.8OCaml #2 20.2420.241,788377  100% 0% 0% 0%
6.8Clean 20.2520.251,448572  0% 0% 0% 100%
7.3Ada 2005 GNAT 21.7621.761,612710  0% 0% 0% 100%
7.4ATS 22.2622.266641179  0% 0% 100% 0%
7.5C# Mono 22.3122.314,996459  0% 0% 0% 100%
7.5F# Mono 22.3722.377,364514  0% 0% 0% 100%
8.4Java 6 -Xint #2 94.6725.1812,952950  93% 94% 92% 98%
8.8Go 6g 8g 26.3226.321,548420  100% 0% 0% 0%
11C++ GNU g++ 32.8132.811,276452  0% 8% 0% 100%
11GNU gcc 33.7733.77684382  0% 100% 0% 0%
20JavaScript TraceMonkey 58.4058.414,428311  0% 100% 0% 0%
22Erlang HiPE 64.4564.4511,692507  100% 0% 0% 0%
30JavaScript V8 90.4590.4524,040311  0% 100% 0% 0%
31Haskell GHC #2 92.6191.1912,256396  1% 100% 1% 1%
44Java 6 -Xint 131.08131.1113,536514  0% 100% 0% 0%
63Scheme PLT #2 187.38187.3812,640419  0% 0% 0% 100%
64Lua 190.68190.701,492329  61% 38% 0% 0%
73Python CPython #5 14 min218.6928,068428  99% 99% 99% 99%
73Lua #3 218.84218.831,480328  0% 0% 0% 100%
74PHP #3 13 min222.4326,3201193  88% 89% 91% 100%
96Perl #3 18 min285.5510,976846  100% 100% 100% 100%
261PHP #2 13 min13 min7,860397  100% 0% 0% 0%
264Ruby JRuby 13 min13 min230,184292  1% 1% 1% 99%
278Python CPython 13 min13 min4,500378  0% 0% 0% 100%
373Perl #2 18 min18 min4,240343  0% 0% 0% 100%
477Perl 23 min23 min3,748355  16% 84% 0% 0%
752Ruby MRI 37 min37 min4,288292  0% 1% 0% 100%
Smalltalk VisualWorks #2 Failed438
interesting alternative programs
1.5GNU gcc #2 4.504.50237,056669
3.8Python CPython #2 11.3211.33473,092233
222Python CPython #4 663.34663.364,608281

 spectral-norm benchmark : Eigenvalue using the power method

diff program output N = 100 with this output file to check your program is correct before contributing.

Each program should calculate the spectral norm of an infinite matrix A, with entries a11=1, a12=1/2, a21=1/3, a13=1/4, a22=1/5, a31=1/6, etc

Each program must implement 4 separate functions / procedures / methods like the C# program.

For more information see challenge #3 in Eric W. Weisstein, "Hundred-Dollar, Hundred-Digit Challenge Problems" and "Spectral Norm".

From MathWorld--A Wolfram Web Resource.
http://mathworld.wolfram.com/Hundred-DollarHundred-DigitChallengeProblems.html
http://mathworld.wolfram.com/SpectralNorm.html

Thanks to Sebastien Loisel for this benchmark.

Revised BSD license