21.2.2 Data-level Parallelism
For some applications, data naturally comes in vector or matrix form. For example, a vector of digitized samples representing an audio waveform over time, or a matrix of pixel colors in a 2D image from a camera. When processing that data, it's common to perform the same sequence of operations on each data element. The example code shown here is computing a vector sum, where each component of one vector is added to the corresponding component of another vector. By replicating the datapath portion of our CPU, we can design special-purpose vector processors capable of performing the same operation on many data elements in parallel. Here we see that the register file and ALU have been replicated and the control signals from decoding the current instruction are shared by all the datapaths. Data is fetched from memory in big blocks (very much like fetching a cache line) and the specified register in each datapath is loaded with one of the words from the block. Similarly eac...