Needs["OpenCLLink`"]
srcf = FileNameJoin[{$OpenCLLinkPath, "SupportFiles", "matrixMul.cl"}]
blockSize = 4;
MatrixMultiply = OpenCLFunctionLoad[File[srcf],
"matrixMul", {{"Float", 2, "Output"}, {"Float", 2,
"Input"}, {"Float", 2, "Input"}, {"Local", "Float"}, {"Local",
"Float"}, _Integer, _Integer}, {blockSize, blockSize},
"Defines" -> {"BLOCK_SIZE" -> blockSize}]
A = RandomChoice[{0, 1}, {8, 8}];
B = RandomChoice[{0, 1}, {8, 8}];
out = OpenCLMemoryAllocate["Float", {8, 8}]
MatrixMultiply[out, A, B, blockSize*blockSize, blockSize*blockSize,8,8]