Models such as DALL-E, Stable Diffusion, Jukebox, and ChatGPT have taken the world by storm by demonstrating surprising abilities to create art, prose, music, and even compute code. AI has clearly reached a tipping point, but we believe it’s just the beginning with this technology. This is why we are thrilled to announce that Vultr will be among the first cloud providers to bring the latest generation of accelerated computing power to our virtual machines and bare metal servers.
The NVIDIA HGX H100 platform, built with NVIDIA H100 Tensor Core GPUs based on the Hopper architecture and NVIDIA networking, joins the current lineup of Vultr’s Cloud GPU offerings, which include NVIDIA A100 Tensor Core, A40, and A16 GPUs. The HGX H100 is a high-performance computing (HPC) system designed specifically for artificial intelligence (AI) and HPC workloads.
The NVIDIA HGX H100 features fourth-generation Tensor Cores and the Transformer Engine with FP8 precision, providing up to 30x faster performance over the prior generation for AI inference with 530B-parameter large language models.
With fourth-generation Tensor Cores and fourth-generation NVLink interconnect technology, the NVIDIA HGX H100 is capable of delivering up to 1.5 exaflops of AI performance. This enables our customers to perform complex workloads with lightning-fast speed and accuracy.
One of the key benefits of the NVIDIA HGX H100 is its ability to handle large, complex models and datasets easily. This makes it ideal for use in the healthcare, finance, and manufacturing industries, where data-intensive tasks are common. Additionally, the H100 can handle multiple workloads simultaneously, making it suitable for use in a variety of applications, including machine learning, deep learning, and data analytics.
To optimize application performance and accelerate the path to production AI deployments with NVIDIA HGX H100, customers can leverage NVIDIA AI Enterprise, a secure, end-to-end, cloud-native suite of AI software. NVIDIA AI Enterprise includes an extensive library of full-stack software, including NVIDIA AI workflows, frameworks, pretrained models, and infrastructure optimization.
Vultr's support for the NVIDIA HGX H100 will allow developers and innovators to access this powerful technology on a pay-as-you-go basis. The platform is highly flexible and can be configured to meet the unique needs of our customers’ workloads, without the need for expensive upfront investments in hardware. This means that they can scale their computing resources up or down as needed, making it a cost-effective solution for organizations of any size.
At Vultr, we are committed to providing our customers with the latest advancements in technology, and the addition of the NVIDIA HGX H100 platform is a testament to this commitment. Soon, Vultr customers will be able to deploy this next-generation GPU on VMs and bare metal.
We encourage you to reserve the NVIDIA HGX H100 GPU now. And, if you’re not already using cloud GPUs, start training with the NVIDIA A100 Tensor Core GPU today by signing up for a Vultr account.