On the Correct Measurement of Application Memory Bandwidth and Memory Access Latency

2019/12/17 by Christian Helm, Kenjiro Taura · 1 voice
Computer Science · Engineering · #Advanced Data Storage Technologies #Parallel Computing and Optimization Techniques #Radiation Effects in Electronics

paper · doi:10.1145/3368474.3368476

openalex publication_date 2019/12/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/29

Abstract

Diagnosing if an application suffers from DRAM contention can be a challenging task. One method is to compare the hardware memory bandwidth limit with the measured memory bandwidth of an application. Another method is based on memory access latency. The latency of a DRAM access in an uncontended state is a hardware characteristic. If an application shows higher DRAM access latency, the increase comes from queuing delays and the application is limited by DRAM bandwidth. Hardware-based measurement of the application's latency and bandwidth can be done with low-overhead and is agnostic of the application's implementation.

Citations

• Introducing a Performance Model for Bandwidth-Limited Loop Kernels

Discussions

• On the Correct Measurement of Application Memory Bandwidth and Access Latency [hn, 1 points, 0 comments]

Related

• Understanding the Interactions of Workloads and DRAM Types: A Comprehensive Experimental Study
• Adaptive-Latency DRAM: Reducing DRAM Latency by Exploiting Timing Margins
• Reducing DRAM Latency at Low Cost by Exploiting Heterogeneity
• Flexible-Latency DRAM: Understanding and Exploiting Latency Variation in Modern DRAM Chips
• Understanding and Improving the Latency of DRAM-Based Memory Systems