计算机科学
带宽(计算)
建筑
缩放比例
电子工程
计算机体系结构
系统体系结构
信号处理
钥匙(锁)
频率标度
嵌入式系统
多路复用
标识
DOI:10.1109/lca.2026.3705360
摘要
High-frequency HBM4 suffers from bandwidth underutilization because each pseudo channel (pCH) has only two bank groups within a selected Stack ID (SID), making BG interleaving insufficient to hide long column-to-column timing constraints. Existing industrial approaches mitigate this problem by adding DQ TSV resources, increasing implementation cost. We propose SPIRE-HBM, a sub-pseudo-channel-based HBM4 architecture that restores seamless BG-interleaved reads without increasing DQ TSVs. SPIRE-HBM partitions each 32-DQ pCH into two 16-DQ sub-pCH data paths across Stack IDs and doubles the burst length to preserve conventional access granularity. The longer burst provides sufficient column-command separation for two-BG interleaving, shifting the limiting column timing from$\text{tCCD}_{\text{L}}$to$\text{tCCD}_{\text{S}}$. Our evaluation shows that SPIRE-HBM improves performance by up to 1.10× for Llama3-8B and 1.15× for synthetic workloads over HBM4. In addition, at 12.8 Gbps, SPIRE-HBM achieves performance comparable to a TSV-doubled HBM4 baseline while retaining the DQ TSV count.
科研通智能强力驱动
Strongly Powered by AbleSci AI