NVIDIA and HP Inc. Collaborate to Enhance Data Processing for AI Development
NVIDIA and HP Inc. have announced a partnership to integrate NVIDIA CUDA-X data processing libraries with HP AI workstation solutions. This integration aims to boost the efficiency of data preparation and processing tasks that are crucial for generative AI development.
Utilizing the NVIDIA CUDA compute platform, CUDA-X libraries are designed to accelerate data processing for various data types such as tables, text, images, and video. One of the key components is the NVIDIA RAPIDS cuDF library, which can significantly speed up the work of data scientists using pandas software by up to 110 times when using an NVIDIA RTX 6000 Ada Generation GPU instead of a CPU-only system, all without the need for any code modifications. These libraries, including RAPIDS cuDF, will be accessible through Z by HP AI Studio on HP AI workstations, providing a comprehensive development solution that enhances data science workflows.
"Pandas is an essential tool for millions of data scientists involved in data processing for generative AI," said Jensen Huang, founder and CEO of NVIDIA. "By accelerating pandas without requiring any code changes, we are taking a significant step forward. Data scientists can now process data much faster, enabling them to handle significantly larger datasets for training generative AI models."
"Data science is the backbone of AI, and developers require fast access to software and systems to support this critical work," stated Enrique Lores, president and CEO of HP Inc. "With the integration of NVIDIA AI software and accelerated GPU compute, HP AI workstations offer a robust solution for our customers."
NVIDIA CUDA-X Enhances Data Science on HP Workstation Solutions
Pandas offers a powerful data structure known as DataFrames, which allows developers to easily manipulate, clean, and analyze tabular data.
The NVIDIA RAPIDS cuDF library accelerates pandas to run on GPUs seamlessly without any code changes, eliminating the reliance on CPUs that can slow down workloads as data sizes increase. RAPIDS cuDF is compatible with third-party libraries and streamlines GPU and CPU workflows, enabling data scientists to develop, test, and deploy models efficiently.
As datasets continue to expand, RTX 6000 Ada Generation GPUs provide 48 GB of memory per GPU to handle large data science and AI workloads on HP Z by workstations. With the capability to accommodate up to four RTX 6000 GPUs, the HP Z8 Fury stands out as one of the most powerful workstations for AI development. The collaboration between HP and NVIDIA allows data scientists to simplify their development process by working on local systems to process even the most demanding generative AI workloads.
Availability
NVIDIA RAPIDS cuDF for accelerated pandas without code changes is set to be available on HP AI workstation solutions featuring NVIDIA RTX and GeForce RTX GPUs this month, with availability on HP AI Studio expected later this year.
