MATLAB-CUDA

  MATLAB平台下的CUDA加速库。运用该库可避免在MEX函数中来回拷贝GPU数据，提高MATLAB运行速率。附件给出一个简单的矩阵乘法的示例程序，该程序在GTX TITAN下运行效率比intel i7 3930k提高了10倍。(CUDA MATLAB platform acceleration libraries. Avoid the use of the library can be copied back and forth GPU MEX function data, improve MATLAB run rate. Annex gives an example of a simple matrix multiplication program that runs under the GTX TITAN 10 times more efficient than the intel i7 3930k.)

MatrixMul demo
..............\cuda 5.0
..............\........\Cfunc.h,410,2014-06-08
..............\........\demo.m,1738,2014-06-08
..............\........\EnD2H.mexw32,137216,2014-06-08
..............\........\EnDestroy.mexw32,55296,2014-06-07
..............\........\EnFree.mexw32,136704,2014-06-08
..............\........\EnH2D.mexw32,137728,2014-06-08
..............\........\EnInitial.mexw32,55808,2014-06-08
..............\........\EnsunDll.dll,30720,2014-06-07
..............\........\EnsunMatrixMul.cu,2330,2014-06-08
..............\........\EnsunMatrixMul.mexw32,139776,2014-06-08
..............\Ensun Kernel Libruary（simple version）.doc,40960,2014-06-07
..............\readme.txt,59,2014-06-08