精选理由
中国团队用14nm工艺做出了520 TFLOPS的AI芯片,全靠架构创新,带宽6.4TB/s,很硬核。
中国发布首款结合软件定义与3D近存计算的AI芯片,在14nm工艺下实现520 TFLOPS算力。该芯片通过架构创新而非工艺微缩,内存带宽达到6.4TB/s。相比传统芯片,它在同等工艺下性能提升显著。这标志着中国在AI芯片架构领域取得关键进展。
原文 · Pandaily
520 TFLOPS at 14nm: China Self-Developed AI Chip Achieves Architecture Breakthrough With Software-Defined Computing
China first AI chip combining software-defined and 3D near-memory computing delivers 520 TFLOPS at 14nm, achieving 6.4TB/s memory bandwidth through architecture innovation instead of process scaling.