Intel unveils new CPUs with up to 96 cores, targeted at AI and HPC workloads

Intel has just launched its new line of Xeon Scalable processors, codenamed Sapphire Rapids. These new processors are aimed at high-performance computing (HPC) and artificial intelligence (AI) workloads, with some models featuring up to 96 cores. This represents a significant increase in performance over the previous generation of Xeon Scalable processors, which had a maximum of 40 cores.

The new processors are based on Intel’s 10nm Enhanced SuperFin technology and come with a range of new features to improve performance, security, and manageability. The processors feature support for PCIe 5.0, DDR5 memory, and Intel’s new Compute Express Link (CXL) technology, which enables faster communication between processors and other components.

In addition to the increased core count and new features, the new processors also come with improved AI and HPC capabilities. They include Intel’s Deep Learning Boost (DL Boost) technology, which is designed to accelerate AI workloads, as well as new vector extensions for HPC workloads.

Intel is positioning the new processors as ideal for use in data centers and cloud computing environments, where AI and HPC workloads are becoming increasingly common. The company claims that the new processors offer up to 1.5 times better performance for AI workloads and up to 2 times better performance for HPC workloads, compared to the previous generation.

Overall, the new Xeon Scalable processors represent a significant step forward in Intel’s efforts to compete in the growing AI and HPC markets. With their increased core count, new features, and improved performance, they are sure to find a home in many data centers and cloud computing environments around the world.