International Conference on Instrumentation, Measurement, Circuits and Systems (ICIMCS 2011)
313 The Research of Acceleration Method of Matrix Operations Based on Embedded Linux
Download citation file:
The paper based on the Linux operating system, Taking the most two common matrix algorithms (matrix multiplication and matrix inverse) into the kernel with the method of module to add system calls. and then, using the crossing compile to take the kernel module to the target application board, and providing two interfaces for the external function in the end.. As is shown in the result: The operation speed and time that loaded the matrix algorithm module into the kernel more faster than in the application layer.