NextIO has released the vCore C200 GPU Consolidation Appliance, a product that brings enterprise-class features to a GPU compute platform. NextIO's vCore C200 delivers serviceability, manageability and flexibility to today’s GPU farms through increased server uptime and reduced GPU over-provisioning.
NextIO's vCORE C200 solves the issue of managing and scaling GPUs by allowing Nvidia Tesla GPUs to be added 'on-demand' to servers as required. By pooling, sharing and dynamically reconfiguring up to eight Tesla GPUs per server, vCore addresses the need for managing GPUs in a dynamic HPC environment. Its modular architecture also allows GPUs to be hot-swapped or updated without disrupting jobs already running on the system.
NextIO's vCore product family offers Nvidia's GPUs for both compute-intensive and data visualisation needs. With the ability to hold eight double-wide GPUs in 4U of space, the vCore C200 delivers more than eight TeraFLOPS of GPU computing power that can be provisioned and managed through a simple-to-use interface. The fully serviceable and redundant design, together with NextIO's vConnect switching technology, allows vCore to map Tesla GPUs to servers dynamically as needed, increasing the GPU utilisation in HPC clusters. The hot-plug capability and serviceable chassis provides quick repair of GPUs and eliminates the need for 'forklift' upgrades to migrate to the newest GPU. In addition, GPUs can be driven by industry-standard resource schedulers to deliver GPU compute power when needed, which makes vCore C200 ideal for medium-to-large clusters.