2403.14123
arxiv.org · 3,999 words · saved by 1 readers
N/A
NOTICE: This arXiv version is an extended version of our paper that has been published in IEEE Micro Journal. AI and Memory Wall Amir Gholami1,2 Zhewei Yao1 Sehoon Kim1 Coleman Hooper1 Michael W. Mahoney1,2,3 Kurt Keutzer1 1 2 3…
related reading
- Transformer Inference Arithmetic | kipply's blogkipp.ly
- Transformer Math 101 | EleutherAI Blogblog.eleuther.ai
- Transformers Inference Optimization Toolset | AstraBlogastralord.github.io
- Can AI scaling continue through 2030? | Epoch AIepoch.ai
- [2205.14135] FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awarenessarxiv.org
- [2510.13786] The Art of Scaling Reinforcement Learning Compute for LLMsarxiv.org
- Memory bandwidth constraints imply economies of scale in AI inference — LessWronglesswrong.com
- AI Chip Architecturesjacobpeake.com
- arxiv.org/pdf/2505.24832arxiv.org
- Data movement bottlenecks to large-scale model training: Scaling past 1e28 FLOP | Epoch AIepoch.ai
- 2305.14314arxiv.org
- Paged Attention from First Principles: A View Inside vLLM – Hamza's Bloghamzaelshafie.bearblog.dev