Radio Free HPC: The Persistence of Memory

Print Friendly, PDF & Email

By Shahin Khan and Dan Olds, Orionx.net

Nothing says HPC like having the performance cake and eating it too.

In this episode, we drill down on what Intel is doing with their cool Optane memory tech, shooting for speeds that remind you of memory, sizes that look like storage, and costs that make it look like a deal, with real byte-addressable persistent memory right inside the server – or block-addressable, if you want. This is a space that was bound to get filled and we’ve been watching the industry’s progress.

The RFHPC team welcomes special guest Kristie Mann to discuss. Really great choice it turns out since she’s Senior Director of Optane DC Persistent Memory Products… at Intel. So yes, the conversation suddenly became pretty real. We’re getting closer to what Shahin called a “Big Wall of Persistent Memory” in a prediction some years ago.

We had a lively chat so no Cyber-Scare from Henry. No Catches this week either, but here’s a mini catch: you get extra points if the title above reminded you of the 1931 Dali painting of the same name. HPC is also about making surreal performance real!