The future of AI computing is shaped by RISC-V processors with multi-core technology and AI capabilities, offering computational efficiency and power.
Sophgo has announced licensing several high-performance processor cores from SiFive, including the SiFive Performance P670 and SiFive Intelligence X280, for developing RISC-V-based AI computing processors.
The SG2380 is a 2.5 GHz, 16-core RISC-V processor that employs the SiFive Performance P670 in a multi-cluster, multi-core setup. It offers superior performance and leverages the computational benefits of RISC-V vector processing. The SG2380 also includes the SiFive Intelligence X280 and a Vector Cooperative Processor Interface Extension (VCIX), integrating AI/ML hardware design from Sophgo. With full RISC-V software toolchain support, this coprocessor is aimed at AI/ML and a range of advanced computing applications.
The P670 combines RISC-V’s unique instruction set architecture with advanced microarchitecture and silicon design from the SiFive team, delivering high computational density and vector processing. Additionally, Sophgo uses the X280, combined with VCIX, as a Neural Processing Unit (NPU), showcasing the partnership’s strengths.
Some of the SG2380 specifications include:
- 16-core SiFive P670 (RV64GCVH) 64-bit RISC-V processor @ up to 2.5GHz with RISC-V Vector v1.0, Vector Crypto
- Cluster configuration – 12x 2.5 GHz performance cores, 4x 1.6 GHz efficiency cores
- Full RISC-V RVA22 proﬁle compliance
- Imagination AXT-16-512 high-mid-performance 3D GPU with support for Vulkan 1.3, OpenGL 3.0, OpenGL ES 3.x/2.0/1.1; 0.5 TFLOPS, 16 Gpixels, and 2 TOPS
- 2D graphic engine
- Video Processing Unit (VPU)
- Up to 4Kp60 10-bit H.265/HEVC, 8-bit H.264/AVC, 8-/10-bit AV1, 8-/10-bit VP9
- No hardware video encoder
The press release highlights that the SiFive P670 is the most powerful, spec-compliant RISC-V commercial-grade processor IP, with SpecINT2k6 exceeding 13/GHz. It supports multi-cluster, multi-core configurations compliant with RVA22, aligning with the latest RISC-V specifications and operating systems for optimal performance and efficiency. The X280 excels in AI applications, offering versatility and scalability for multi-core, multi-cluster setups, with a track record of successful chip deployments. The development tools and software are aligned with the OpenXLA framework, catering to current market demands and future AI application trends.
For more information, click here.