Abstract
With the emergence of large capacity persistent memory modules, such as Intel® Optane™ DC Persistent Memory Modules, applications now can access this high bandwidth, low latency storage tier. Newer usages are emerging from HPC, IOT and AI segments where you need larger memory capacity but may not need to persist the data. In this session, we will show you different options for converting large capacity persistent memory to volatile memory and with the minimum development effort. We will explore two different libraries: 1.libmemkind, a general purpose allocation library that supports different "kinds" of memory & 2. libvmemcache, a dedicated caching library that implements a more optimal memory management scheme.