简介：
Linpack是我们常用的CPU性能测试程序。它通过计算双精度线性方程组的求解来测试CPU的运算能力。Intel MKL提供一个优化版本的Intel® Optimized LINPACK Benchmark，通过运行这个程序，我们可以方便进行CPU的基准性能测试。
Intel® Optimized LINPACK Benchmark是根据LINPACK 1000 benchmark优化后的程序。程序根据用户指定的参数生成一个线性的方程组，通过方程的求解时间与计算量，来计算CPU的浮点性能。
配置解析：
配置文件：
Sample Intel(R) Optimized LINPACK Benchmark data file (lininput_xeon32)Intel(R) Optimized LINPACK Benchmark data1# number of tests 测试的方程组数量5511 # number of equations (problem sizes) 方程的数量5511 # leading dimensions 矩阵的主维度10 # number of times to run a test (trials) 运行的次数4# alignment values (in KBytes) 内存地址对齐值
测试过程：
linpack_11.1.2/benchmarks/linpack/xlinpack_xeon64 <config_file>
结果解析：
Intel(R) Optimized LINPACK Benchmark dataCurrent date/time: Wed Jun 19 16:30:36 CPU frequency: 0.795 GHzNumber of CPUs: 2Number of cores: 20Number of threads: 40Parameters are set to:Number of tests: 1Number of equations to solve (problem size) : 5511Leading dimension of array : 5511Number of trials to run : 10Data alignment value (in Kbytes) : 4Maximum memory requested that can be used=243083284, at the size=5511=================== Timing linear equation system solver ===================Size LDA Align. Time(s) GFlops ResidualResidual(norm) Check5511 5511 41.46376.2963 1.765377e-11 2.038689e-02 pass5511 5511 41.25588.9343 1.765377e-11 2.038689e-02 pass5511 5511 41.41878.7118 1.765377e-11 2.038689e-02 pass5511 5511 41.28686.7912 1.765377e-11 2.038689e-02 pass5511 5511 41.43377.9163 1.765377e-11 2.038689e-02 pass5511 5511 41.23690.3218 1.765377e-11 2.038689e-02 pass5511 5511 41.52273.3548 1.765377e-11 2.038689e-02 pass5511 5511 41.34383.1036 1.765377e-11 2.038689e-02 pass5511 5511 41.40279.6433 1.765377e-11 2.038689e-02 pass5511 5511 41.44777.1696 1.765377e-11 2.038689e-02 passPerformance Summary (GFlops)Size LDA Align. Average Maximal5511 5511 4 81.2243 90.3218Residual checks PASSEDEnd of tests
关键指标：
AverageGFlops(每秒10亿次的浮点运算数)
