Overview
The course aims to explore the impact of cache management on the performance of I/O intensive applications, specifically focusing on direct cache access (DCA) and Data Direct I/O technology (DDIO) in Intel processors. The goal is to optimize DDIO-enabled systems for I/O intensive applications, demonstrating potential latency reductions of up to ~30% at 100 Gbps. The course covers topics such as the workings of DDIO, LLC ways used by DDIO, tuning DDIO for performance, and the implications of DDIO on processing time. The intended audience includes system designers, developers, and individuals interested in optimizing systems for multi-hundred-gigabit networks.
Syllabus
Intro
Direct Cache Access (DCA)
Intel Data Direct I/O (DDIO)
Pressure from these trends
What happens at 200 Gbps?
How does DDIO work?
LLC ways used by DDIO
How does DDIO perform?
Reducing #Descriptors is Not Sufficient! (1/2)
IIO LLC WAYS Register
Impact of Tuning DDIO
Is Tuning DDIO Enough?
What about Current Systems?
Using Our Knowledge for 200 Gbps
Our Key Findings (1/2)
Impact of Processing Time
Conclusion
Taught by
USENIX