Benchmarks report

Generated on 2011-07-13, 05:26 AM

CPU: Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz

Total memory: 8173548 kB

Input file: /home/spiros/git/auto-numerical-bench/app-benchmarks/autobench/files/python/cblastests.in
# Testing different implementations of CBLAS

# cblas-reference is a wrapper for blas: test this with different blas implementations
reference-eigen sci-libs/cblas-reference-20110218 blas:eigen
reference-openblas sci-libs/cblas-reference-20110218 blas:openblas-threads
reference-acml sci-libs/cblas-reference-20110218 blas:acml64-gfortran

# Other implementations are not wrappers
# Test them using gcc-4.6.1 and enabling vectorization
gsl gsl-1.15-r1 CC=gcc-4.6.1 CFLAGS="-O3 -pipe -march=native -msse3 -msse4.1 -msse4.2"
goto sci-libs/gotoblas2-1.13 CC=gcc-4.6.1 CFLAGS="-O3 -pipe -march=native -msse3 -msse4.1 -msse4.2" USE=incblas TARGET=NEHALEM
openblas sci-libs/openblas CC=gcc-4.6.1 CFLAGS="-O3 -pipe -march=native -msse3 -msse4.1 -msse4.2" USE=incblas TARGET=NEHALEM
Logs: /var/log/benchmarks/cblas_2011-07-13

Summary

y = a*x + y

Apply Givens rotation

Matrix-Vector multiply

Transposed Matrix-Vector multiply (A' * x)

Rank-1 update

Triangular system solution

Matrix-Matrix multiply

Symmetric Rank-n update

Triangular system solution with n right hand side vectors