Intel Xeon Processors Accelerate GenAI Workloads with Aible

In a significant advancement for the enterprise AI landscape, Intel and Aible have joined forces to offer groundbreaking solutions for running advanced generative AI (GenAI) and retrieval-augmented generation (RAG) use cases. This collaboration, which leverages the robust capabilities of Intel® Xeon® CPUs, promises to lower costs, enhance efficiency, and embed intelligence into enterprise applications.

Aible, known for its serverless generative AI and augmented analytics solutions, has optimized its technology to run seamlessly on multiple generations of Intel Xeon processors. This partnership encompasses engineering optimizations and a comprehensive benchmarking program, designed to demonstrate the superior performance of CPUs in handling sophisticated AI workloads. The result is a scalable and efficient AI solution that capitalizes on high-performing hardware to address the pressing needs of modern enterprises.

Mishali Naik, a senior principal engineer at Intel’s Data Center and AI Group, underscored the importance of this collaboration. “Customers are looking for efficient, enterprise-grade solutions to harness the power of AI,” Naik said. “Our collaboration with Aible shows how we’re closely working with the industry to deliver innovation in AI and lowering the barrier to entry for many customers to run the latest GenAI workloads using Intel Xeon processors.”

Aible’s solutions illustrate the potential of CPUs to significantly enhance performance across a variety of AI tasks, from running language models to executing RAG. The company’s technology employs an efficient serverless end-to-end approach, activating resources only when user requests are made. For instance, the vector database and language model power up briefly to process queries, thus reducing the total cost of ownership (TCO).

Traditionally, RAG implementations have relied on GPUs and accelerators for their parallel processing prowess. However, Aible’s innovative serverless technique, combined with Intel Xeon Scalable processors, demonstrates that CPUs alone can efficiently power RAG use cases. Performance data reveals that multiple generations of Intel Xeon processors are more than capable of handling these demanding workloads.

The implications of this development are profound for enterprise customers. Aible’s benchmark analysis indicates that clients can achieve up to a 55-fold cost saving when running RAG models on their CPU-based serverless solutions. This remarkable cost efficiency stems from Aible’s CPU-exclusive strategy, which eliminates the need for expensive GPU-based infrastructures and allows shared resources to be utilized securely across multiple customers.

Intel’s collaboration with Aible extends beyond hardware. The two companies have worked closely to optimize AI workloads on Xeon processors, with significant improvements realized through code optimization for AVX-512. These strategic software enhancements have led to notable performance gains and improved throughput, highlighting the crucial role of software in maximizing hardware efficiency.

The combination of RAG models with Intel Xeon processors, facilitated by platforms like Aible, opens up new possibilities for applications in natural language processing (NLP), recommendation systems, decision support systems, and content generation. This collaboration, which began with the launch of Intel’s 4th Gen Xeon processors, has continuously evolved to optimize AI workloads, code, and libraries, ensuring that Aible’s product offerings remain at the forefront of technological innovation.

Intel and Aible are set to showcase their solutions at the Amazon Web Services Summit in Washington, D.C., on June 26 and 27. Aible’s solutions, which run on AWS Lambda, are available in the AWS Marketplace, making them accessible to a broader audience of enterprise customers.

The collaboration between Intel and Aible exemplifies the transformative potential of strategic partnerships in the AI industry, promising a future where advanced AI workloads are more accessible, efficient, and cost-effective.

Read the full report (Aible.com) | 30 Days to AI Value: Development Best Practices from Intel and Aible (Intel.com) | Impact from AI in 30 Days (Aible Case Study) | Intel AI Analytics Toolkit

Tags: Artificial Intelligence, Data Center & HPC

Category: Intel

This entry was posted on Thursday, June 27th, 2024 at 4:54 pm. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.