In a compelling keynote at the Databricks Data + AI Summit 2024, NVIDIA Founder and CEO Jensen Huang highlighted the transformative power of accelerated computing in data processing and analytics. In a conversation with Databrick’s co-founder and CEO Ali Ghodsi, Huang described enterprise data as a “gold mine” for businesses, underscoring the challenges and opportunities in harnessing this valuable resource.
Databricks leverages NVIDIA’s full stack to accelerate generative AI applications
Highlighting a major development, Huang and Ghodsi announced the integration of NVIDIA’s accelerated computing technology with Databricks Photon, a powerful data processing engine. This collaboration aims to improve the performance and cost-efficiency of Databricks SQL, leveraging NVIDIA’s capabilities in accelerated computing and generative AI.
“This is a big announcement,” Huang noted, noting the critical role of accelerated computing and generative AI in today’s technology landscape. “NVIDIA and Databricks combine our expertise to deliver these improvements to enterprises.”
The integration, which has been in the works for five years, involves a series of specialized libraries from NVIDIA to accelerate Photon, enabling faster and more energy-efficient data processing.
“We are excited to work with NVIDIA to leverage GPU acceleration on the Photon engine,” said Ghodsi. “This will improve core data processing capabilities and allow them to run on NVIDIA GPUs.”
Creating generative AI factories with NVIDIA NIM
In addition, NVIDIA and Databricks announced that Databricks’ open-source model DBRX is now available as an NVIDIA NIM microservice, accessible through the NVIDIA API catalog. These microservices provide fully optimized, pre-built containers for deployment across environments, simplifying the integration of generative AI models into applications.
“Creating these endpoints is complicated,” Huang noted. “We optimized everything into a microservice that runs both in the cloud and on-premises, increasing developer productivity by standardizing the integration of AI models.”
Launched in March, DBRX leverages Databricks’ extensive tools and techniques and is trained using the NVIDIA DGX Cloud platform. Companies can customize DBRX with their data to create specific models or use it to develop a combination of expert models.
Huang emphasized the importance of accelerating data processing and adopting generative AI, urging companies to engage with these rapidly advancing technologies. “Generative AI is growing exponentially – you don’t want to be left behind,” he advised.
Join the conversation
Summit attendees were encouraged to join sessions and engage with NVIDIA experts to explore the future of AI and data intelligence. Key sessions include:
- “Development and Deployment of Generative AI with NVIDIA” at 12:30 PM PT
- “Architecture Analysis for ETL Processing: CPU vs. GPU” at 4:30 PM PT
- “Spark RAPIDS ML: GPU Accelerated Distributed ML in Spark Clusters” at 1:30 PM PT
For more information, visit the NVIDIA Blog.
Image source: Shutterstock
. . .
Labels