spectral-norm benchmark N=5,500

Each chart bar shows how many times slower, one ↓ spectral-norm 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.

     sortsortsort
  ×   Program Source Code CPU secs Elapsed secs Memory KB Code B ≈ CPU Load
1.0Fortran Intel #2 7.867.87784513  0% 2% 100% 0%
1.0C gcc #5 7.861.981,704569  99% 100% 99% 99%
1.0C gcc #4 7.881.981,5681139  99% 99% 99% 99%
1.0Fortran Intel #3 7.911.995,384638  100% 99% 100% 100%
1.0Fortran Intel 7.921.995,440568  99% 100% 100% 100%
1.0C++ g++ #5 7.952.002,0201044  100% 100% 100% 100%
1.0C++ g++ #6 7.962.002,0241044  100% 100% 100% 100%
1.2Ada 2005 GNAT #4 9.362.373,8682762  99% 99% 99% 99%
1.7C++ g++ #8 13.183.311,5481278  100% 100% 100% 100%
2.0Go #3 15.703.951,024536  99% 99% 99% 100%
2.0C gcc #3 15.723.941,708463  100% 100% 100% 100%
2.0Pascal Free Pascal 15.7615.778423  1% 1% 0% 100%
2.0Go #2 15.814.161,096668  96% 95% 96% 94%
2.0Lisp SBCL #2 15.944.1317,772906  96% 97% 97% 97%
2.0Ada 2005 GNAT 15.9515.951,628710  0% 0% 100% 1%
2.0Scala 16.0715.9828,140404  100% 1% 1% 1%
2.0Haskell GHC #4 16.084.151,392984  95% 97% 99% 97%
2.1Ada 2005 GNAT #3 16.134.083,9561702  100% 99% 98% 99%
2.1Dart #5 16.1616.0120,220518  1% 9% 92% 1%
2.1Dart 16.2416.0920,220489  1% 100% 0% 1%
2.1C gcc 16.2816.28876383  0% 0% 100% 0%
2.1Scala #5 16.404.4034,436693  92% 93% 97% 92%
2.1Java  #2 16.534.3026,476950  99% 96% 95% 96%
2.2Java  16.9916.9126,444514  100% 0% 1% 1%
2.3OCaml #3 18.145.103,192938  93% 93% 95% 93%
2.3F# Mono #3 18.276.5443,412720  66% 69% 73% 72%
2.4OCaml #2 18.8518.861,656377  0% 1% 1% 100%
2.6Go 20.8020.801,024411  1% 0% 100% 1%
2.9Racket #3 22.945.9921,612627  96% 95% 95% 97%
3.0Lisp SBCL 23.3923.4010,176625  0% 0% 100% 0%
3.0Lisp SBCL #3 23.936.1720,072883  97% 97% 97% 97%
3.6Racket #2 28.5028.4919,344532  0% 0% 1% 100%
3.7C++ g++ 29.1729.181,224452  1% 1% 2% 100%
3.8C# Mono 29.5429.5539,124459  100% 0% 1% 0%
4.0Racket 31.2831.2721,136446  0% 0% 1% 100%
4.0C# Mono #2 31.688.1738,4121063  97% 97% 98% 97%
4.0F# Mono #2 31.749.9042,832852  77% 77% 81% 87%
5.3Clojure #6 41.6411.3553,388808  93% 93% 91% 90%
5.4Clojure #8 42.4011.3753,944918  96% 92% 92% 94%
5.7Clojure #7 44.9412.1653,128762  91% 91% 91% 98%
9.1Erlang HiPE #2 71.4518.1920,904747  98% 98% 98% 99%
14Haskell GHC #2 112.1837.849,880403  66% 65% 100% 66%
15Erlang HiPE 114.35114.3831,488507  0% 100% 0% 1%
27Perl #5 211.13211.195,468340  1% 100% 0% 0%
27Perl #4 211.8054.416,644551  97% 98% 96% 98%
29Perl 226.90226.983,936333  1% 1% 0% 100%
38Ruby JRuby #4 295.36288.54649,160326  26% 28% 23% 28%
52Ruby JRuby 6 min6 min647,980292  25% 24% 27% 27%
56PHP #3 7 min110.1814,1361193  99% 99% 100% 99%
56PHP #2 7 min7 min4,688397  0% 100% 1% 1%
71Ruby #4 9 min9 min5,812326  1% 0% 65% 36%
80Ruby 10 min10 min6,712292  0% 1% 24% 77%
82Ruby #3 10 min168.1425,624828  96% 97% 97% 97%
120Python 3 #6 15 min15 min4,980328  97% 1% 4% 0%
127Ruby JRuby #2 16 min258.84656,984776  97% 98% 97% 98%
128Python 3 #5 16 min252.9230,320437  99% 100% 99% 99%
134Perl #2 17 min17 min3,088343  1% 0% 100% 1%
135Perl #3 17 min267.098,192846  100% 99% 99% 100%
Rust Make Error898
Scala #2 Failed720
Scala #3 Failed982
Scala #4 Failed1006
"wrong" (different) algorithm / less comparable programs
0.3C gcc #2 2.322.32237,796669
1.1Python 3 #2 8.528.53479,784233
1.1C++ g++ #7 8.542.152,0401283
2.2C++ g++ #2 17.284.341,6001330
9.6Python 3 #3 75.4975.59957,216379

 spectral-norm benchmark : Eigenvalue using the power method

You can write your own program for this task and contribute to the benchmarks game by following these general instructions.

More specifically:

diff program output N = 100 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.

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

  Home   Conclusions   License   Play