#AVX-512

News stories tagged with #AVX-512

Technology Update: New Intel Architectures and Attractive Lenovo Business Laptops in Focus

This report compiles current developments from the world of hardware and software, ranging from new CPU architectures to current hardware offers. The focus is on the progress of Intel Nova Lake regarding AVX-512 and a concrete offer for powerful Lenovo Business Laptops. The summary provides readers with both technical insights and purchasing opportunities in the business segment.

Widely Covered

Optimizing Linux: How CPU Instructions and Driver Development Boost Open Source Performance

Canonical is testing x86-64-v3 package builds for Ubuntu 26.10 to leverage modern CPU instructions like AVX/AVX2 and enhance system performance. Concurrently, Google has achieved significant performance gains in Linux RAID through a revised AVX-512 implementation. These advancements occur alongside efforts to improve hardware compatibility, particularly for graphics drivers between FreeBSD and modern Intel/AMD systems.

Intel Shifts Strategy: SambaNova Deal Falls Through, Xeon Partnership and Granite Rapids Updates Announced

Intel has scrapped its planned acquisition of AI chip company SambaNova and instead agreed to a strategic partnership based on Xeon CPUs. The company also released technical details on its Granite Rapids-WS Xeon 600 processors, including reduced boost speeds due to AVX-512 and AMX instructions, and issued an updated microcode for the Granite Rapids D series to fix hardware-specific issues. These developments reflect Intel's evolving strategy in the competitive AI hardware market while addressing ongoing technical challenges.

Widely Covered

AMD Ryzen 10000 'Olympic Ridge' with Zen 6 and 2 nm: Delayed to 2027, but with AVX-512 and Up to 24 Cores

AMD has delayed the release of its next-generation Ryzen processors, codenamed 'Olympic Ridge' with Zen 6 architecture, to 2027 due to challenges with DDR5 memory supply and supply chains. The processors will use the AM5 socket, feature up to 24 cores, 48 MB of L3 cache per CCD, and support AVX-512 for consumer applications. Meanwhile, AMD is also developing the Ryzen 500-Series 'Medusa Point' APU, which will include a mobile-optimized RDNA 4m integrated GPU with FSR 4 support and LPDDR6 memory, with the 'Medusa Halo' variant expected to use the more powerful RDNA 5 architecture.