Liqid pours on the GPUs

The company maximizes GPU flexibility for solid, fast M&E storage.

Liqid, a software company delivering data center composability, presented its adaptive GPU platform and what it claims to be the fastest NVMe storage solution for the Media & Entertainment market at NAB 2022.

Liqid uses a software-defined approach to resource management while maximizing resource utilization, allowing GPUs to work at peak performance. Solutions based on Liqid’s Matrix software enable users to deploy previously impossible amounts of GPU and NVMe storage in minutes, versus having to physically alter hardware configurations. Such resources also can be portioned in exact amounts via a simple GUI.

“The most advanced GPU servers and disk-based storage solutions are limited by their form factor, trapped and underutilized,” said Nader Soudah, vice president of Liqid’s Global Channel. “Liqid Matrix-based solutions remove those physical limitations with software that delivers significant cost savings and time-to-value for producers, but just as importantly, gives time back to creators and crews, whether they’re rendering dailies on set, building the next VFX marvel at a workstation, or working magic in an editing bay.”

Liqid’s Matrix CDI software enables users to maximize GPU flexibility by configuring up to 16 GPUs to compute nodes to meet performance demands with the turnkey Liqid ThinkTank system, whereby under-utilized GPUs can be moved to other compute nodes in real time to satisfy shifting workload requirements. Customers can also get up to four million IOPS per SSD and over 24GB/s throughput with the company’s LQD4500 NVMe SSD, aka the HoneyBadger.