From: Paul C. Leopardi (paul.leopardi_at_[hidden])
Date: 2007-01-24 07:05:30
Dear Gunter,
I have appended the results of the test below, with my processor details.
I'm running SUSE Linux 10.1.
Best, Paul
On Wednesday 24 January 2007 10:11, Gunter Winkler wrote:
...
> can you also try
>
> #include <boost/numeric/ublas/vector_of_vector.hpp>
>
> generalized_vector_of_vector< double, row_major, vector< X > >
>
> with X = compressed_vector<double>, coordinate_vector<double>,
> mapped_vector<double>
Dense: ublas::matrix
Matrix addition test:
Dim = 4, CPU = 0.00; Inner product = 0; Fill = 0.5
Dim = 8, CPU = 0.00; Inner product = 0; Fill = 0.5
Dim = 16, CPU = 0.00; Inner product = 0; Fill = 0.5
Dim = 32, CPU = 0.01; Inner product = 0; Fill = 0.5
Dim = 64, CPU = 0.04; Inner product = 0; Fill = 0.5
Dim = 128, CPU = 0.14; Inner product = 0; Fill = 0.5
Dim = 256, CPU = 0.59; Inner product = 0; Fill = 0.5
GVV Mapped: ublas::generalized_vector_of_vector<vector<mapped_vector<double> >
>
Matrix addition test:
Dim = 4, CPU = 0.00; Inner product = 0; Fill = 0.5
Dim = 8, CPU = 0.01; Inner product = 0; Fill = 0.5
Dim = 16, CPU = 0.07; Inner product = 0; Fill = 0.5
Dim = 32, CPU = 0.29; Inner product = 0; Fill = 0.5
Dim = 64, CPU = 1.16; Inner product = 0; Fill = 0.5
Dim = 128, CPU = 6.80; Inner product = 0; Fill = 0.5
Dim = 256, CPU = 27.00; Inner product = 0; Fill = 0.5
GVV Compressed:
ublas::generalized_vector_of_vector<vector<compressed_vector<double> > >
Matrix addition test:
Dim = 4, CPU = 0.01; Inner product = 0; Fill = 0.5
Dim = 8, CPU = 0.02; Inner product = 0; Fill = 0.5
Dim = 16, CPU = 0.05; Inner product = 0; Fill = 0.5
Dim = 32, CPU = 0.17; Inner product = 0; Fill = 0.5
Dim = 64, CPU = 0.54; Inner product = 0; Fill = 0.5
Dim = 128, CPU = 1.83; Inner product = 0; Fill = 0.5
Dim = 256, CPU = 7.29; Inner product = 0; Fill = 0.5
GVV Coordinate:
ublas::generalized_vector_of_vector<vector<coordinate_vector<double> > >
Matrix addition test:
Dim = 4, CPU = 0.01; Inner product = 0; Fill = 0.5
Dim = 8, CPU = 0.02; Inner product = 0; Fill = 0.5
Dim = 16, CPU = 0.08; Inner product = 0; Fill = 0.5
Dim = 32, CPU = 0.22; Inner product = 0; Fill = 0.5
Dim = 64, CPU = 0.69; Inner product = 0; Fill = 0.5
Dim = 128, CPU = 2.37; Inner product = 0; Fill = 0.5
Dim = 256, CPU = 9.60; Inner product = 0; Fill = 0.5
Mapped: ublas::mapped_matrix
Matrix addition test:
Dim = 4, CPU = 0.01; Inner product = 0; Fill = 0.5
Dim = 8, CPU = 0.03; Inner product = 0; Fill = 0.5
Dim = 16, CPU = 0.09; Inner product = 0; Fill = 0.5
Dim = 32, CPU = 0.35; Inner product = 0; Fill = 0.5
Dim = 64, CPU = 1.36; Inner product = 0; Fill = 0.5
Dim = 128, CPU = 7.80; Inner product = 0; Fill = 0.5
Dim = 256, CPU = 33.03; Inner product = 0; Fill = 0.5
Compressed: ublas::compressed_matrix
Matrix addition test:
Dim = 4, CPU = 0.01; Inner product = 0; Fill = 0.5
Dim = 8, CPU = 0.02; Inner product = 0; Fill = 0.5
Dim = 16, CPU = 0.04; Inner product = 0; Fill = 0.5
Dim = 32, CPU = 0.36; Inner product = 0; Fill = 0.5
Dim = 64, CPU = 4.83; Inner product = 0; Fill = 0.5
Dim = 128, CPU = 113.33; Inner product = 0; Fill = 0.5
Dim = 256, CPU = 2160.00; Inner product = 0; Fill = 0.5
Coordinate: ublas::coordinate_matrix
Matrix addition test:
Dim = 4, CPU = 0.01; Inner product = 0; Fill = 0.5
Dim = 8, CPU = 0.06; Inner product = 0; Fill = 0.5
Dim = 16, CPU = 0.54; Inner product = 0; Fill = 0.5
Dim = 32, CPU = 8.10; Inner product = 0; Fill = 0.5
Dim = 64, CPU = 220.00; Inner product = 0; Fill = 0.5
Dim = 128, CPU = 4410.00; Inner product = 0; Fill = 0.5
Dim = 256, CPU = 132100.00; Inner product = 0; Fill = 0.5
leopardi_at_linfinit:~/src/test_add> cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 4
model name : AMD Athlon(tm) 64 Processor 3200+
stepping : 8
cpu MHz : 800.000
cache size : 1024 KB
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca
cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext lm 3dnowext 3dnow
bogomips : 1609.06
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp