Unverified50% confidenceFactExact time
Under a 45nm process, a single 32-bit floating-point multiplication consumes approximately 3.7pJ of energy, reading 32-bit data from off-chip DRAM consumes approximately 640pJ, and on-chip SRAM access requires approximately 5pJ.
1
Sources
50%
Confidence
Long-term
Relevance
8/19/2026
First Seen
Sources
Related Claims
UnverifiedH100的内存带宽约3.35TB/s，GPT-4规模模型单次生成一个Token需读取约1TB参数数据，理论上每秒最多生成约3个Token75% similarUnverifiedWSE-3的片上SRAM容量达到44GB，SRAM访问延迟约1ns，比HBM约100ns低10-50倍74% similarUnverified以 LLaMA-2 70B 为例，每次完整推理需从存储加载约 140GB 权重数据，以 5GB/s 的 SSD 读速估算仅数据加载时间就约需 28 秒72% similarUnverifiedIf the quantized model weights are 150 GB, reading the weights once at a bandwidth of 273 GB/s takes approximately 0.55 seconds, giving a theoretical single-token generation speed of approximately 1.8 tokens/s (excluding computation latency and cross-node communication overhead).71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/775582API
curl https://kongchang.com/api/v1/knowledge/claims/775582MCP
get_claim(id=775582)