SK Hynix AI-Specific Computing Memory Solution: From AiM Device to Heterogeneous AiMX-xPU System for Comprehensive LLM Inference

*Recap Accelerator-in-Memory (AiM) & AiMX *System Extensions of AiMX Card for Datacenter *AiM & AiMX for On-device AI *Design Choices for Future AiM/AiMX *Conclusion

Saved in:
Bibliographic Details
Main Authors: Kim, Guhyun, Kim, Jinkwon, Kim, Nahsung, Shin, Woojae, Won, Jongsoon, Joo, Hyunha, Choi, Haerang, An, Byeongju, Shin, Gyeongcheol, Yun, Dayeon, Kim, Jeongbin, Kim, Changhyun, Kim, Ilkon, Park, Jaehan, Song, Yosub, Yang, Byeongsu, Lee, Hyeongdeok, Park, Seungyeong, Lee, Wonjun, Kim, Seonghun, Park, Yonghoon, Jung, Yousub, Park, Gi-Ho, Lim, Euicheol
Format: Conference Proceeding
Language:English
Subjects:
Arithmetic
Artificial intelligence
Computational modeling
Memory architecture
Transformers
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:*Recap Accelerator-in-Memory (AiM) & AiMX *System Extensions of AiMX Card for Datacenter *AiM & AiMX for On-device AI *Design Choices for Future AiM/AiMX *Conclusion
ISSN:2573-2048
DOI:10.1109/HCS61935.2024.10664793