benchmark octave with atlas and vecLib

[Gavin Jin]

I just ran Octave2.m, a benchmark test file I obtained from
http://www.sciviews.org/other/benchmark.htm, on the same powerMac G4 
dual 1.25GHz
with 2GB of RAM with the following two version and library combinations:

1. octave-atlas I installed with fink
GNU Octave, version 2.1.35 (powerpc-apple-darwin6.3).


The results are:



   Octave Benchmark 2
   ==================
Number of times each test is run__________________________: 3

   I. Matrix calculation
   ---------------------
Creation, transp., deformation of a 1500x1500 matrix (sec): 6.91675
800x800 normal distributed random matrix ^1000______ (sec): 0.250428
Sorting of 2,000,000 random values__________________ (sec): 8.45625
700x700 cross-product matrix (b = a' * a)___________ (sec): 0.581424
Linear regression over a 600x600 matrix (c = a \ b') (sec): 1.62466
                  ------------------------------------------------------
                Trimmed geom. mean (2 extremes eliminated): 1.86947

   II. Matrix functions
   --------------------
FFT over 800,000 random values______________________ (sec): 0.701015
Eigenvalues of a 320x320 random matrix______________ (sec): 3.43814
Determinant of a 650x650 random matrix______________ (sec): 2.09488
Cholesky decomposition of a 900x900 matrix__________ (sec): 0.309856
Inverse of a 400x400 random matrix__________________ (sec): 0.894866
                  ------------------------------------------------------
                Trimmed geom. mean (2 extremes eliminated): 1.09534

   III. Programmation
   ------------------
750,000 Fibonacci numbers calculation (vector calc)_ (sec): 1.02013
Creation of a 2250x2250 Hilbert matrix (matrix calc) (sec): 1.40126
error: `gcd2' undefined near line 234 column 9
error: evaluating index expression near line 234, column 9
error: evaluating assignment expression near line 234, column 7
error: evaluating for command near line 230, column 1
error: near line 237 of file 
`/Volumes/development/code/code/matlabOrOctave/benchmark/Octave2.m'




2. Octave 2.1.50 with vecLib and fftw that I obtained from
http://homepage.mac.com/chibaf/math/macosx/install/


Here are the results:



   Octave Benchmark 2
   ==================
Number of times each test is run__________________________: 3

   I. Matrix calculation
   ---------------------
Creation, transp., deformation of a 1500x1500 matrix (sec): 2.479
800x800 normal distributed random matrix ^1000______ (sec): 0.2555
Sorting of 2,000,000 random values__________________ (sec): 9.244
700x700 cross-product matrix (b = a' * a)___________ (sec): 0.5975
Linear regression over a 600x600 matrix (c = a \ b') (sec): 0.2609
                  ------------------------------------------------------
                Trimmed geom. mean (2 extremes eliminated): 0.7284

   II. Matrix functions
   --------------------
FFT over 800,000 random values______________________ (sec): 1.142
Eigenvalues of a 320x320 random matrix______________ (sec): 2.393
Determinant of a 650x650 random matrix______________ (sec): 0.2972
Cholesky decomposition of a 900x900 matrix__________ (sec): 0.3238
Inverse of a 400x400 random matrix__________________ (sec): 0.1864
                  ------------------------------------------------------
                Trimmed geom. mean (2 extremes eliminated): 0.479

   III. Programmation
   ------------------
750,000 Fibonacci numbers calculation (vector calc)_ (sec): 1.047
Creation of a 2250x2250 Hilbert matrix (matrix calc) (sec): 1.473
error: `gcd2' undefined near line 234 column 9
error: evaluating assignment expression near line 234, column 7
error: evaluating for command near line 230, column 1
error: near line 237 of file 
`/Volumes/development/code/code/matlabOrOctave/benchmark/Octave2.m'



I'm not sure why we're seeing impressive performance with vecLib only 
on I.A, I.E, II.C, and II.E, but not other tests. 



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------