Atlas Library installation

[sangamesh.bc]

Hi All,

I have installed the ATLAS (Mathematics) Library in my home directory on a amd64 quadcore based linux cluster.

The coomands used for this are:
make
make install arch=Linux_HAMMER64SSE2_4

What are steps to be performed after the installation? i.e. which environment variables, PATH variables have to be set?
Can any one listout the these steps, so that ATLAS can be used by other compilers/applications?

Another copy is already installed in /opt. But its having some problem. So I installed new one in my home dir. This lib has to be used by some other applications. How to make that application to use this lib which is installed by me instead of the existing /opt copy.

regards,
Sangamesh

[newtolinuxguy]

i just did the install myself. i actually used as little variables/options as possible and it worked (make check passed ok). however, i do have questions about [make time]. my question is at the bottom of this post.

bunzip2 -c atlas3.8.0.tar.bz2 | tar xfm - # create SRCdir
cd ATLAS # enter SRCdir
mkdir MyObj # create BLDdir
cd MyObj # enter BLDdir
../configure
make build # tune & build lib
make check # sanity check correct answer
make time # check if lib is fast
make install # copy libs to install dir

my question is:
hi,

below is the output from [make time] after installing ATLAS. i'm not
sure how to interpret the % that are negative or greater than 100%.
has anyone seen this before?

i have/use:
dell dimension E520/intel core duo 2
ubuntu 7.04
atlas 3.8.0

The times labeled Reference are for ATLAS as installed by the authors.
NAMING ABBREVIATIONS:
kSelMM : selected matmul kernel (may be hand-tuned)
kGenMM : generated matmul kernel
kMM_NT : worst no-copy kernel
kMM_TN : best no-copy kernel
BIG_MM : large GEMM timing (usually N=1600); estimate of asymptotic
peak
kMV_N : NoTranspose matvec kernel
kMV_T : Transpose matvec kernel
kGER : GER (rank-1 update) kernel
Kernel routines are not called by the user directly, and their
performance is often somewhat different than the total
algorithm (eg, dGER perf may differ from dkGER)

Reference clock rate=2394Mhz, new rate=1862Mhz
Refrenc : % of clock rate achieved by reference install
Present : % of clock rate achieved by present ATLAS install

single precision double precision
********************************
*******************************
real complex real
complex
--------------- --------------- ---------------
---------------
Benchmark Refrenc Pres

ent Refrenc Present Refrenc Present Refrenc Present
========= ======= ======= ======= ======= ======= ======= =======
=======
kSelMM 535.0 -251.6 525.4 -225.6 311.5 319.6
298.0 295.4
kGenMM 175.5 -251.6 175.5 -225.6 160.5 165.6
165.4 168.6
kMM_NT 145.2 -199.2 149.3 -191.9 135.3 126.7
132.3 134.1
kMM_TN 163.2 -204.9 161.1 -207.1 148.7 143.9
146.0 157.9
BIG_MM 510.1 171.8 504.0 172.3 307.7 319.7
293.0 312.2
kMV_N 113.5 26.9 216.9 0.0 58.9 66.4
97.4 104.5
kMV_T 89.9 0.0 94.6 107.4 47.2 53.7
74.1 86.6
kGER 154.2 0.0 119.4 107.4 29.1 31.3
46.8 58.9

my question is also posted at:
interpreting timing of ATLAS - sci.math.num-analysis | Google Groups