linpack benchmark

[Dr Beco <rcb@beco.cc>]

Dear linuxers,


How do I run a linpack benchmark in my computer?

I was able to compile this old code (parts from 1978!)
http://www.netlib.org/benchmark/linpackc
and run, but I'm afraid it may be innacurate or do not have the
capacity to measure multiple cores.

So I tried to find a more recent code, till I, for my surprise, find
out the debian package:

$ dpkg -s hpcc
Package: hpcc
Status: install ok installed
Priority: extra
Section: science
Installed-Size: 1568
Maintainer: Debian Science Maintainers
<debian-science-maintainers@lists.alioth.debian.org>
Architecture: amd64
Version: 1.4.1-2
Depends: libatlas3gf-base, libc6 (>= 2.7), libopenmpi1.3, mpi-default-bin
Description: HPC Challenge benchmark
 The High Performance Computing (HPC) Challenge benchmark runs a suite
 of 7 tests that measure the performance of CPU, memory and network for
 HPC clusters.  Amongst others, it includes the High-Performance LINPACK
 (HPL) benchmark, used by the Top500 ranking (http://www.top500.org/).
Homepage: http://icl.cs.utk.edu/hpcc/


But there is no manual page, nor I can find instructions to run.
Simple running hpcc gives an error:

$ hpcc
HPL WARNING from process # 0, on line 313 of function HPL_pdinfo:
>>> cannot open file hpccinf.txt <<<


Is there a guideline? I just want to see the FLOPS, nothing to complex
I presume.
And, of course, to configure the test to use full capabilities of the
machine (an intel i5). I suspect hpccinf.txt is for that, isn't it?


BTW, from the compiled program, I got most of the times:


$ ./linpakc
Rolled Double Precision Linpack

Rolled Double Precision Linpack

     norm. resid      resid           machep         x[0]-1        x[n-1]-1
       1.7        7.41628980e-14  2.22044605e-16 -1.49880108e-14 -1.89848137e-14
    times are reported for matrices of order   100
      dgefa      dgesl      total       kflops     unit      ratio
 times for array with leading dimension of  201
       0.00       0.00       0.00        inf       0.00       0.00
       0.00       0.00       0.00        inf       0.00       0.00
       0.00       0.00       0.00        inf       0.00       0.00
       0.00       0.00       0.00     858333       0.00       0.01
 times for array with leading dimension of 200
       0.00       0.00       0.00        inf       0.00       0.00
       0.00       0.00       0.00        inf       0.00       0.00
       0.00       0.00       0.00        inf       0.00       0.00
       0.00       0.00       0.00    1716667       0.00       0.01
Rolled Double  Precision 858333 Kflops ; 10 Reps


but sometimes the number is:
Rolled Double  Precision 1716667 Kflops ; 10 Reps
and I even got a negative value once in a while, which raised
suspicious about the code.

$ ./linpakc
     norm. resid      resid           machep         x[0]-1        x[n-1]-1
       1.7        7.41628980e-14  2.22044605e-16 -1.49880108e-14 -1.89848137e-14
Rolled Double  Precision -2147483648 Kflops ; 10 Reps


compiled with:

gcc linpakc.c -o linpakc -DDP -DROLL -O4 -lm


Thanks,
Beco.




-- 
Dr Beco
A.I. researcher

--> . <--

"Look again at that dot. That's here. That's home. That's us. On it
everyone you love, everyone you know, everyone you ever heard of,
every human being who ever was, lived out their lives. The aggregate
of our joy and suffering, thousands of confident religions,
ideologies, and economic doctrines, every hunter and forager, every
hero and coward, every creator and destroyer of civilization, every
king and peasant, every young couple in love, every mother and father,
hopeful child, inventor and explorer, every teacher of morals, every
corrupt politician, every "superstar", every "supreme leader", every
saint and sinner in the history of our species lived there - on a mote
of dust suspended in a sunbeam." (Carl Sagan, 1934-1996)