VIFMM: Custom RISC-V Vector Instructions with INT-FP Mixed-Precision Computing for Accelerating LLM Inference

Kai Gu, Jian Fang, Guangda Zhang, Bingxi Pei, Tianbo Zhang. VIFMM: Custom RISC-V Vector Instructions with INT-FP Mixed-Precision Computing for Accelerating LLM Inference. In 2025 IEEE International Conference on High Performance Computing and Communications, HPCC 2025, Exeter, United Kingdom, August 13-15, 2025. pages 314-321, IEEE, 2025. [doi]

Authors

Kai Gu

This author has not been identified. Look up 'Kai Gu' in Google

Jian Fang

This author has not been identified. Look up 'Jian Fang' in Google

Guangda Zhang

This author has not been identified. Look up 'Guangda Zhang' in Google

Bingxi Pei

This author has not been identified. Look up 'Bingxi Pei' in Google

Tianbo Zhang

This author has not been identified. Look up 'Tianbo Zhang' in Google