Delay-line memory

CPU speed grows faster than memory speed

The memory wall problem highlights the growing gap between CPU and memory speeds, which can hinder the overall performance of computing systems. As CPUs become faster, the slower memory speeds can become a bottleneck, limiting the potential for further advancements in processing power and efficiency.

Example

In early computing systems, delay-line memory had capacities of a few thousand bits with recirculation times measured in microseconds. As CPUs evolved, their processing speeds increased significantly, while memory speeds lagged behind, creating a bottleneck in data transfer and processing.

Understanding the memory wall problem is crucial for designing efficient computing systems that can fully utilize the advancements in CPU technology without being hindered by slower memory speeds.

Related concepts

Von Neumann architecture

CPU must fetch both data and instructions from memory

Overdrawn at the Memory Bank

Overdrawn at the Memory Bank was shot on videotape due to budget constraints

Flashbulb memory

Flashbulb memories are vivid but not always accurate

CPU cache

L1/L2 cache hierarchy reduces global memory latency

GQA reduces KV-cache memory by the group factor

GQA reduces KV-cache memory by dividing storage by the number of groups

Triton auto-tunes BLOCK_SIZE: different sizes optimize for different hardware

Triton auto-tunes BLOCK_SIZE for hardware efficiency, optimizing memory access patterns and computational throughput

One email a day: 5 concepts + the 5 stories that matter →

Swipe through 100 ML concepts daily

Open TickerNews