Tailslayer: C++ Library for Reducing RAM Tail Latency from DRAM Refresh Stalls
By
hasheddan
Front-window bakery material. Catches the eye, delivers the goods.
Summary
Tailslayer is a C++ library designed to reduce tail latency in RAM reads caused by DRAM refresh stalls. It works by replicating data across multiple independent DRAM channels with uncorrelated refresh schedules, using undocumented channel scrambling offsets that work on AMD, Intel, and Graviton processors. The library implements hedged reads across all replicas, allowing work to proceed with whichever result responds first. The article provides technical details about the library's implementation and usage.
Key quotes
· 4 pulledTailslayer is a C++ library that reduces tail latency in RAM reads caused by DRAM refresh stalls.
It replicates data across multiple, independent DRAM channels with uncorrelated refresh schedules, using (undocumented!) channel scrambling offsets that works on AMD, Intel, and Graviton.
Once the request comes in, Tailslayer issues hedged reads across all replicas, allowing the work to be performed on whichever result responds first.
The library code is available in hedged_reader.cpp and the example using the library can be found in tailslayer_example.cpp.
You might also wanna read
Assembly Language Programming: The Enduring Power of Low-Level Code
This article explores the enduring relevance and unique advantages of assembly language programming, using the example of Chris Sawyer's dev
The Evolution of Multi-Core Programming: From Specialized Technique to Default Requirement
This article explores the paradigm shift in programming from single-core to multi-core processing as a fundamental requirement rather than a
Technical Analysis of Ultrassembler's High-Speed RISC-V Assembly Performance
Ultrassembler is a high-performance RISC-V assembler library developed as part of the Chata signal processing project. The article explains
Why Average LLM Use Is Likely Destroying Value in Software Development
The author argues that, contrary to prevailing hype, the average use of Large Language Models (LLMs) is likely destroying value rather than
How AI Accelerated Prototyping: From Idea to Tangible in Record Time
The author reflects on how AI has transformed their prototyping workflow. Previously, the biggest bottleneck was the time needed to scaffold
GitLab 19.0 launches with Secrets Manager, agentic workflows, and self-hosted AI models
GitLab 19.0 has been released, positioning itself as an intelligent orchestration platform for DevSecOps. The release includes expanded secr
bit.ly·1d ago