Arm releases new Neoverse V1 and Neoverse N2 Platforms & CPUs
Arm announced that Arm Neoverse for cloud-to-edge infrastructure will launch two new platforms, Neoverse V1 and Neoverse N2. Among them, Neoverse V1 is the first platform of the V series, and Neoverse N2 is the second generation of the N series.
Compared with Neoverse N1, Neoverse V1 has a 50% improvement in a single thread. Arm said that for applications with higher CPU performance and bandwidth requirements, Neoverse V1 is very suitable and can provide its chip partners with better design flexibility.
At the same time, Neoverse V1 supports Scalable Vector Extensions (SVE) and supports two 256-bit width vectors, allowing single instruction stream multiple data stream (SIMD) integer, bfloat16, and floating-point instructions to be executed on a wider vector unit, bringing huge application potential to markets such as high-performance cloud, high-performance computing, and machine learning.
Arm has just launched the Armv9 architecture to meet user requirements for a new generation of high-performance processors, which paved the way for the launch of Neoverse N2. Neoverse N2 is the first platform based on Armv9 architecture, with improvements in safety, energy efficiency, and performance.
Compared with Neoverse N1, Neoverse N2 increases single-thread performance by 40% while maintaining the same level of power and area efficiency. The performance of NGINX is 1.3 times higher than the previous generation. Arm said that Neoverse N2 can provide excellent performance per thread and industry-leading performance per watt, thereby reducing overall costs for users. Neoverse N2 is the first platform to support SVE2, which can significantly improve the performance of the cloud to the edge but also brings ease of programming and portability.