Question \#1 (Memory Hierarchy) (30 points) A programmer developing image processing algorithms found that her program frequently calls transpose procedure on 20482048 double precision matrices. A sample code is illustrated below. She is using the C programming language that follows the row major order convention, that is, the elements of a matrix are placed in memory as follows IN[0,0],IN[0,1],IN[0,2],IN[0,3], etc. Her task is to improve performance of the matrix transpose operation by utilizing optimization techniques and her knowledge of the underlying architecture. The processor has a 64KB fully associative first level data cache with 64-byte cache blocks. Assume the level 1 data cache has write-allocate policy on write misses. Assume write-backs do not consume clock cycles. for(i=0;i