Is RISC-V ready for High Performance Computing? An evaluation of the Sophon SG2044

2025-08-19

AI Summary

Empirical evaluation of RISC-V processors in high-performance computing (HPC) remains scarce, particularly for domestically developed many-core chips. Method: This work presents the first systematic assessment of the 64-core RISC-V processor Sophon SG2044 for HPC workloads, leveraging the RISC-V Vector Extension (RVV) v1.0 and an optimized memory subsystem to build a multicore parallel benchmarking framework. Contribution/Results: Evaluation across representative compute-intensive HPC benchmarks shows that SG2044 achieves up to a 4.91× speedup over its predecessor. Under 64-core concurrency, it demonstrates significantly improved floating-point throughput and memory bandwidth efficiency. Its performance gap relative to state-of-the-art x86 and ARM HPC platforms is substantially narrowed, with several workloads reaching comparable levels. This study provides critical empirical evidence and architectural optimization insights for RISC-V’s adoption in HPC.

Technology Category

• Machine Learning: Hardware-aware ML
• Humans and AI: Other Foundations of Human Computation & AI
• Computer Vision: Learning & Optimization for CV

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Large-scale security measurements

Abstract

The pace of RISC-V adoption continues to grow rapidly, yet for the successes enjoyed in areas such as embedded computing, RISC-V is yet to gain ubiquity in High Performance Computing (HPC). The Sophon SG2044 is SOPHGO's next generation 64-core high performance CPU that has been designed for workstation and server grade workloads. Building upon the SG2042, subsystems that were a bottleneck in the previous generation have been upgraded. In this paper we undertake the first performance study of the SG2044 for HPC. Comparing against the SG2042 and other architectures, we find that the SG2044 is most advantageous when running at higher core counts, delivering up to 4.91 greater performance than the SG2042 over 64-cores. Two of the most important upgrades in the SG2044 are support for RVV v1.0 and an enhanced memory subsystem. This results in the SG2044 significantly closing the performance gap with other architectures, especially for compute-bound workloads.

Problem

• Evaluating RISC-V's readiness for high performance computing workloads
• Assessing Sophon SG2044 CPU performance compared to previous generation
• Analyzing impact of RVV v1.0 and memory subsystem upgrades

Innovation

• Upgraded memory subsystem for performance
• Support for RVV v1.0 vector extension
• 64-core RISC-V architecture for HPC