Brendan Gregg

FEBRUARY 28, 2023

My personal opinion is that I don't see a widespread need for more capacity given horizontal scaling and servers that can already exceed 1 Tbyte of DRAM; bandwidth is also helpful, but I'd be concerned about the increased latency for adding a hop to more memory. Or even on a plane.

Performance 110

Performance Latency Cache Virtualization 110

Brendan Gregg

FEBRUARY 28, 2023

My personal opinion is that I don't see a widespread need for more capacity given horizontal scaling and servers that can already exceed 1 Tbyte of DRAM; bandwidth is also helpful, but I'd be concerned about the increased latency for adding a hop to more memory. Or even on a plane. It was a great privilege.

Performance 52

Performance Latency Cache Virtualization 52

Adrian Cockcroft

JANUARY 29, 2023

The mean and percentile measurements hide this structure, but the rest of this post will show how the structure can be measured and analyzed so that you can figure out a useful model of your system, understand what is driving the long tail of latencies and come up with better SLAs and measures of capacity.

Lambda 98

Lambda Latency Cache C++ 98

The Netflix TechBlog

NOVEMBER 3, 2022

We introduce a caching mechanism in the API gateway layer, allowing us to offload processing from singleton leader elected controllers without giving up strict data consistency and guarantees clients observe. We started seeing increased response latencies and leader servers running at dangerously high utilization.

Cache 224

Cache Latency Traffic Systems 224

Brendan Gregg

AUGUST 18, 2019

For example, iostat(1), or a monitoring agent, may tell you your average disk latency, but not the distribution of this latency. For smaller environments, it can be of more use helping eliminate latency outliers. Block I/O latency as a histogram. This traces block I/O, and shows latency as a power-of-2 histogram.

Latency 68

Latency C++ Cache Programming 68

Sutter's Mill

FEBRUARY 13, 2017

Tue-Thu Apr 25-27: High-Performance and Low-Latency C++ (Stockholm). On April 25-27, I’ll be in Stockholm (Kista) giving a three-day seminar on “High-Performance and Low-Latency C++.” If you’re interested in attending, please check out the links, and I look forward to meeting and re-meeting many of you there.

Latency 51

Latency C++ Hardware Performance 51

Adrian Cockcroft

JANUARY 20, 2023

Most of the top supercomputers are similar to Frontier, they use AMD or Intel CPUs, with GPU accelerators, and Cray Slingshot or Infiniband networks in a Dragonfly+ configuration. The four categories still make sense: kernel managed network sockets, user mode message passing libraries, coherent memory interfaces, and on-chip communication.

Architecture 52

Architecture Latency Benchmarking AWS 52