Now QCT is collaborating with NVIDIA, joining the NVIDIA AI Enterprise software platform with some of their most powerful GPU servers. NVIDIA AI Enterprise, highlighted at the blockbuster GTC conference last March in San Jose, is an end-to-end, cloud-native software platform that accelerates data science pipelines and streamlines development and deployment of production-grade GenAI applications.

NVIDIA of course is foremost known for its market-dominant GPUs powering AI workloads, but the company also places an enormous emphasis on software that enables the implementation of HPC-class AI deployments at scale, including the NVIDIA CUDA® GPU programming platform and NVIDIA AI Enterprise.

The NVIDIA-QCT collaboration enables customers to deploy AI solutions across a range of infrastructures, including virtualized environments and private clouds in corporate data centers. Together, the NVIDIA AI Enterprise software suite and QCT’s NVIDIA-certified systems create a scalable platform that empowers diverse AI applications, accelerating deep learning and machine learning across various industries.

By leveraging NVIDIA NIM microservices, which is a component of NVIDIA AI Enterprise, enterprises can seamlessly deploy microservices to accelerate the deployment of foundation models across the data center while ensuring data security.

The QCT-NVIDIA collaboration also incorporates NVIDIA AI Enterprise with virtualization software provided by VMware by Broadcom. Offering easily integrated AI into an existing VMware infrastructure, it allows enterprises to run AI workloads on an existing virtualized infrastructure. In addition, its enhanced resource management utilizing GPU sharing allows multiple virtual machines to access the same physical GPU resource.

Moving forward, QCT is also developing VMware Private AI Foundation with NVIDIA to help enterprises address their privacy, choice, cost, performance and compliance concerns. VMware Private AI Foundation with NVIDIA is a joint GenAI platform that integrates VMware and NVIDIA software to enable enterprises to run RAG workflows, fine-tune and customize LLM models, and run inference workloads in their own data centers.

To help enterprises build their modern AI infrastructure, QCT offers powerful NVIDIA- and VMware-certified GPU servers with best-fit hardware configurations tailored for diverse AI needs. Let’s explore their capabilities:

QuantaGrid D54U-3U (Next Generation- QuantaGrid D75E-4U):

The QuantaGrid D54U-3U is a versatile server designed to handle a wide range of AI workloads. It offers exceptional flexibility to support various PCIe-based accelerators. This server accommodates up to 4 dual-width GPUs or up to 8 single-width GPUs, making it ideal for a wide range of GPU applications. Its next-generation model, the QuantaGrid D75E-4U, maintains this flexible architecture while delivering even greater GPU computing power with up to 8 dual-width GPU cards, including NVIDIA H200 NVL GPUs, for advanced performance.

QuantaGrid D74H-7U:

The QuantaGrid D74H-7U is a cutting-edge server designed to accelerate AI training with massive datasets and large-scale AI models. Equipped with 8 NVIDIA H100/H200 SXM5 GPU modules on an NVIDIA HGX™ baseboard, it delivers exceptional performance for even the most demanding AI and HPC workloads.

Source: InsideHPC