While it still has a long way to go before catching up with AWS and Azure on adoption, IBM Corp. is already pulling ahead it comes to the breadth of hardware options available in its public cloud. The technology giant this morning announced that it’s become the first major provider to add support for Nvidia Corp.’s latest-generation Tesla M60 GPUs.

Debuted at VMworld last August, the chip series fits two beefy graphic accelerators on a dual-slot card with 16 gigabytes of GDDRR5 memory. Nvidia says that the unit can support up to 32 concurrent users when deployed in conjunction with its GRID 2.0 optimization software for virtual desktop environments. In total, it’s possible to run as many as 128 instances on a single server, which the GPU maker claims is twice the maximum limit of the optimization engine’s previous release.

As a result, organizations with a lot of virtual desktops to support end up having to rent less hardware from IBM and are thereby able to noticeably cut their operating expenses. The company has named M+W Group GmbH, a top German construction company with more than 6,000 users, as one of the first customers that plan on migrating its new M60-powered machines. Big Blue’s public cloud is also used by firms like MapD Technologies Inc., which harnesses the graphical prowess of GPUs to speed up business intelligence workloads.

The recently funded startup claims its namesake relational store can render data visualizations up to 50 times faster than traditional alternatives designed to use CPUs. And IBM is also marketing its GPU-supported machines to organizations developing artificial intelligence software, which in turn benefits from the density of Nvidia chips. More cores per processor means more nodes can be included in a neural network.

Image via Pixabay