Vultr is the first cloud provider to offer virtualization of the NVIDIA A100 Tensor Core GPU for AI workloads
Vultr®, a leading independent provider of cloud infrastructure, today announced that Vultr Talon, powered by NVIDIA GPUs and NVIDIA AI Enterprise software, is now available in beta. A breakthrough cloud-based platform, Vultr Talon offers affordable accelerated computing by enabling GPU sharing, so multiple workloads can efficiently run on a single NVIDIA GPU.
Vultr is the first cloud provider to offer virtualization of NVIDIA A100 Tensor Core GPUs to enable GPU sharing. With the launch today, Vultr is introducing a set of virtual machine (VM) plans featuring the NVIDIA A100, starting at just $90 per month, or $0.13 per hour.
Cloud Platform Breakthrough Enabling AI Workloads at a Fraction of the Cost
At the core of Vultr Talon is a state-of-the-art NVIDIA GPU virtualization platform built on NVIDIA’s vGPU software. Rather than attaching entire physical GPUs to VMs, Vultr instead attaches just a fraction in the form of a virtual GPU (vGPU). These vGPUs are powered by the NVIDIA AI Enterprise software suite, which includes NVIDIA vGPU software and is optimized for remotely running AI workloads and high-performance data analytics.
To a customer’s machine, a vGPU looks just like a physical GPU. Each vGPU has its own dedicated memory that is a portion of the underlying card’s memory. The vGPU has access to a corresponding portion of the physical GPU’s computational power. For Vultr plans with at least 10GB of GPU RAM, NVIDIA’s Multi-Instance GPU (MIG) technology is used to provide guaranteed QoS, fully isolated GPU high-bandwidth memory cache, and dedicated compute cores to tenants.
High-end GPUs typically cost thousands of dollars per month. While this expense is often justifiable for the largest enterprise workloads, for many businesses and developers, the cost of even a single GPU can be prohibitive to getting started, experimenting, or for running applications in development and testing environments. Even enterprises with substantial IT budgets may end up wasting significant amounts of money, provisioning more GPU capacity than needed, or simply deciding to avoid using GPUs at all.
“At Vultr, we pride ourselves on making high-performance cloud infrastructure affordable for everyone. With Vultr Talon, we have turned the GPU delivery model upside down. Because of our breakthrough cloud platform, AI developers and data scientists can provision exactly as much NVIDIA GPU processing as they actually need, at prices they can afford,” said J.J. Kardwell, CEO of Vultr’s parent company, Constant.
Fractions of an NVIDIA A100, Starting at Just $90 Per Month, or $0.13 Per Hour
The NVIDIA A100 Tensor Core GPU delivers unprecedented acceleration for deep learning, high-performance computing (HPC), and data analytics. Combined with the NVIDIA AI Enterprise software suite, optimized to leverage the benefits of the underlying architecture, NVIDIA A100 accelerates all major deep learning and data analytics frameworks like TensorFlow and over 700 HPC applications via NVIDIA NGC catalog.
Vultr worked closely with NVIDIA to create its Vultr Talon offering, starting with the NVIDIA A100 Tensor Core GPU.
“There’s no one size fits all when it comes to customer workloads, and provisioning the right size acceleration for your workload and maximizing utilization is critical for cloud cost optimization,” said Matthew McGrigg, director of global business development for cloud partners at NVIDIA. “Vultr’s highly accessible platform makes it easy to provision NVIDIA GPU resources with great granularity.”
Starting at just $90 per month, or $0.13 per hour, the initial Vultr Talon plans featuring the virtualized NVIDIA A100 and NVIDIA AI Enterprise software are perfectly suited for machine learning inference and model-building workloads as well as for applications such as natural language processing, voice recognition and computer vision.
Bare Metal GPU for Large Workloads
For customers who wish to run large workloads that require multiple physical GPUs, Vultr is also offering Bare Metal servers with four NVIDIA A100 GPUs and dual 24-core Intel Xeon CPUs.
Ambitious Expansion Ahead
Today’s beta launch is just the beginning for Vultr Talon with NVIDIA GPUs, with initial capacity of the NVIDIA A100 in New Jersey. Vultr will be adding global inventory for NVIDIA A100, A40, and A16 GPUs in the weeks ahead, to better support additional regions and a wider variety of use cases.
Vultr partners such as cloud orchestration platform Cycle.io are enthusiastic about Vultr Talon and Bare Metal GPU products.
“Through our partnership with Vultr, we’re thrilled to not only be able to offer a variety of GPU resources to organizations, but do so in an affordable, and flexible, manner – enough to meet any use case. The ability to mix and match vGPUs and bare metal is game changing for our users,” said Jake Warner, CEO of Cycle.io.
Businesses, developers, and data scientists interested in trying Vultr Talon can provision instances through Vultr’s control panel.