Hybrid cloud data management firm Cloudian Inc. today announced the availability of its new PyTorch connector with Remote Direct Memory Access support that delivers erformance improvements for artificial intelligence and machine learning workloads.

A PyTorch connector is a software interface that allows for data transfer between the popular machine learning framework PyTorch and external storage systems or data sources. A connector helps streamline data loading and improve performance during model training and inference workflows.

Cloudian’s new PyTorch connector is built on Nvidia Corp.’s GPUDirect Storage technology and optimized for Nvidia Spectrum-X networking infrastructure. The connector delivers a 74% increase in data processing performance while reducing processor utilization by 43%, representing a significant advancement in AI workflow acceleration.

In testing using TorchBench, an open-source PyTorch performance measurement tool, the new RDMA-enabled connector processed 52,000 images per second compared to 30,000 images per second using the default Amazon Web Services Inc.’s S3 connector, a substantial performance gain that directly translates to faster model training and reduced computational costs for AI practitioners.

“This represents a fundamental breakthrough in how AI workloads access and process data through advanced Nvidia networking acceleration,” said Neil Stobart, chief technology officer at Cloudian. “By leveraging Nvidia GPUDirect Storage technology to eliminate traditional network bottlenecks, we’re enabling data scientists and AI engineers to supercharge their workflows while reducing infrastructure costs through direct GPU-to-storage communication.”

The connector also benefits from the inclusion of RDMA that allows it to bypass traditional CPU-intensive network protocols, enabling direct memory-to-memory data transfers between Cloudian storage systems and GPU-accelerated AI frameworks running on Nvidia network infrastructure, including Nvidia Spectrum-X Ethernet switches and Nvidia ConnectX SuperNICs.

Cloudian says the advancement will prove particularly significant for PyTorch users leveraging Nvidia accelerated computing. They represent a substantial portion of the machine learning community, including researchers at major technology companies, academic institutions and AI-focused startups.

The PyTorch ecosystem serves millions of developers worldwide, from individual researchers to large-scale enterprise AI operations utilizing Nvidia accelerated computing infrastructure. The new PyTorch connector can assist organizations implementing computer vision, natural language processing and deep learning applications with reduced training times and lower computational overhead delivered by the Nvidia GPUDirect Storage connector.

The new Cloudian PyTorch connector is now available for evaluation, allowing organizations to assess the performance benefits within their Nvidia-accelerated AI environments.

Image: SiliconANGLE/Reve