The new offerings aim to enhance built-in AI acceleration and software optimizations within processors used in data center and edge solutions.
Intel has extended its hardware portfolio for AI and analytics, with the launch of its 3rd gen Intel Xeon Scalable processors, updated persistent memory, and SSD solutions.
The Xeon Scalable is the “first mainstream server processor” with built-in bfloat16 support, according to Intel. “This delivers up to a 1.93X increase in training performance, accelerating time to model deployment and making more frequent re-training practical to keep AI systems up to date,” said Lisa Spelman, Intel vice president and general manager.
“The ability to rapidly deploy AI and data analytics is essential for today’s businesses. We remain committed to enhancing built-in AI acceleration and software optimizations within the processor that powers the world’s data center and edge solutions,” she added.
Intel also announced the Stratix 10 NX FPGA and two specialized hardware accelerators optimized for AI deployment. With this, Intel is almost able to offer the full hardware stack to data centers, but it is missing a crucial piece.
That missing piece is a GPU to compete with Nvidia. In the announcement video, Intel highlighted the GPU as “in development”, but did not provide a timeline on when it was expected to launch.
The GPU is seen as the most important piece for many AI-driven tasks, including self-driving cars. Nvidia is far ahead of all competition in this area, and Intel may struggle to break into the market because of its lateness, as we saw with mobile.