Andes RISC-V Superscalar Multicore A(X)45MP and Vector Processor NX27V Upgrade Their Spec. and Performance

Facebook
Twitter
LinkedIn

SAN JOSE, CA – December 02, 2021 – Andes Technology Corporation (TWSE: 6533; SIN: US03420C2089; ISIN: US03420C1099), a leading supplier of high efficiency, low-power 32/64-bit RISC-V processor cores and Founding Premier member of RISC-V International, today announces its superscalar multicore AndesCore 45MP family and the first commercial RISC-V vector processor IP, AndesCore NX27V, upgrade their spec and performance.

The 8-stage superscalar multiprocessor A(X)45MP was announced one year ago. It supports up to four cores and is equipped with DSP, single/double precision FPU (floating-point unit), and Linux-capable Memory Management Unit. Compared with the previous versions, the upgraded 32-bit A45MP and 64-bit AX45MP deliver up to 3x memory bandwidth while raising the floating-point performance by over 20% as measured by Whetstone benchmark. The latency for Level-1 Cache miss and Level-2 Cache hit is reduced by half and it leads to the outstanding 3.4 SPECint2006/GHz performance. In addition, the upgrade also includes RISC-V trace interface and debug spec.

The NX27V, supporting the RISC-V Vector Extension (RVV) spec v1.0-rc1, is upgraded with full configurations of 128-bit to 512-bit VLEN/SIMD/MEM. For vector data types, the NX27V now implements FP16 to FP64 and Int8 to Int64 as well as Andes-enhanced BF16 and Int4 for optimized AI data representations. The NX27V contains a scalar unit and an Out-of-Order Vector Processing Unit (VPU) with a dedicated interface called Streaming Port to efficiently exchange a large amount of data between NX27V registers and an external hardware engine. The NX27V comes with standard development tools and optimized RVV compute libraries. OpenCL™ with integrated LLVM compiler is available to enable parallel programming on heterogeneous architecture using multiple NX27V processors and a host processor such as the AX45MP. With the all 512-bit configuration, the NX27V can achieve over 98x speedup comparing with pure C program and 66% higher performance for MobileNet-v1 benchmark than the all 256-bit configuration. The NX27V targets the applications with large volumes of data such as AI, AR/VR, computer vision, cryptography, and multimedia.

“The 45-series families are welcomed by the market since introduced last year. They are used in diversified applications, ranging from highend MCU, video processing, WiFi 7, 5G base station, AI accelerators, to enterprise-grade storage devices. We are excited that the newly upgraded A(X)45MP with enhanced memory subsystem and optimized FPU can deliver prominent performance to address a wider range of applications. The industrial leading NX27V vector processor just got another award, the EDA & IP Product Award from EE Times (Nov. 16). It has been adopted by nearly 10 customer projects, targeting cloud accelerators with manycore architecture. In this release, NX27V provides a wider range of configurations to cover a variety of performance/area choices,” said Andes President and CTO Dr. Charlie Su. “Together with the complete software development environment and libraries support, A(X)45MP and NX27V are ready to serve more high-performance applications from the edge to the cloud.”

A(X)45MP and NX27V are available for licensing now. Please visit http://www.andestech.com/ for details or contact Andes sales at sales@andestech.com for more information.

About Andes Technology
Sixteen years in business and a Founding Premier member of RISC-V International, Andes is a leading supplier of high-performance/low-power 32/64-bit embedded processor IP solutions, and the driving force in taking RISC-V mainstream. Andes’ fifth-generation AndeStar™ architecture adopted the RISC-V as the base. Its V5 RISC-V CPU families range from tiny 32-bit cores to advanced 64-bit cores with DSP, FPU, Vector, Linux, superscalar, and/or multicore capabilities. The annual volume of Andes-Embedded SoCs has exceeded 2 billion since 2020 and continues to rise. At the end of Q3 2021, the cumulative volume of Andes-Embedded™ SoCs has reached 9 billion. For more information, please visit https://www.andestech.com. Follow Andes on LinkedIn, Twitter, Facebook, and YouTube!