Over the past few decades, the criteria for assessing computer performance have been relatively simple and clear-cut. For CPUs, the focus has been on clock speed and the number of cores; GPUs compete based on their floating-point operation capabilities. Meanwhile, memory and storage devices have primarily emphasized frequency and sequential transfer rates. However, with the rapid evolution of modern processor computing power, a critical challenge has arisen: guaranteeing that data is precisely delivered to the intended destination precisely when it is required. As computing power has surged far ahead of data transmission technology, the expenses linked to data movement—in terms of latency, bandwidth, and energy consumption—have soared. Consequently, memory performance has emerged as a bottleneck in modern computing.