How to Manage High-Bandwidth Memory Automatically
Rathish Das, Kunal Agrawal, Michael A. Bender, Jonathan W. Berry, Benjamin Moseley, Cynthia A. Phillips · 2020
This paper develops an algorithmic foundation for automated management of the multilevel-memory systems common to new supercomputers. In particular, the High-Bandwidth Memory (HBM) of these systems has a similar latency to that of DRAM and a smaller capacity, but it has much larger bandwidth. Systems equipped with HBM do not fit in classic memory-hierarchy models due to HBM's atypical characteristics.