| knu = require("knu") -- https://github.com/knu/hs-knu | |
| -- Switch between Karabiner-Elements profiles as Barrier enters a different host | |
| do | |
| -- Configure Barrier (https://github.com/debauchee/barrier) to output log to ~/Library/Logs/barrier.log | |
| local logFile = os.getenv("HOME") .. "/Library/Logs/barrier.log" | |
| local lineNo = 1 | |
| local host = nil | |
| local defaultProfile = "Default" | |
| local profileForHost = function (host) |
This is a short post that explains how to write a high-performance matrix multiplication program on modern processors. In this tutorial I will use a single core of the Skylake-client CPU with AVX2, but the principles in this post also apply to other processors with different instruction sets (such as AVX512).
Matrix multiplication is a mathematical operation that defines the product of
