The Memory Wall

Cross-cuts: Memory
last updated 2026-06-22
Von Neumann BottleneckHBM (High-Bandwidth Memory)HBM / CoWoS BottleneckPrefill vs Decode (LLM inference phases)Processing-in-Memory (PIM)In-Memory ComputingThe Memory …

The growing gap between how fast processors can compute and how fast memory can feed them. Coined by Wulf and McKee (1995): if compute throughput grows faster than memory bandwidth and latency, then sooner or later every workload becomes memory-bound, and adding more compute stops helping.

By the numbers (the divergence)

Why it binds now

AI inference made the wall the dominant constraint. A modern accelerator can do far more multiply-accumulates per second than its memory can supply operands for, so on memory-bandwidth-bound workloads the expensive silicon sits idle waiting on data. The economics compound it: HBM (High-Bandwidth Memory) (stacked DRAM with through-silicon vias) is the industry’s answer to push the wall back, but it is supply-constrained and expensive (see Hbm Bottleneck and HBM / CoWoS Bottleneck).

Where it bites hardest

Architectural responses

Two families. Move memory closer to compute (Near-Memory Compute, HBM), or do the compute inside the memory so the data never moves (In-Memory Computing, Processing-in-Memory (PIM), Charge-Domain Compute). Most of the post-Moore accelerator landscape is a response to this one wall.

Connected ideas

Related concepts

Frontier questions