gubatron

Things to remember when compiling/linking C/C++ software

by Angel Leon. March 17, 2015;

Last update on December 14, 2023

Updated on February 27, 2023

Updated August 29, 2019.

nadavrot

High-Performance Matrix Multiplication

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).

Intro

Matrix multiplication is a mathematical operation that defines the product of

wakita

SDK = xcrun -sdk macosx

all: compute.metallib compute

compute.metallib: Compute.metal
	# Metal intermediate representation (.air)
	$(SDK) metal -c -Wall -Wextra -std=osx-metal2.0 -o /tmp/Compute.air $^
	# Metal library (.metallib)
	$(SDK) metallib -o $@ /tmp/Compute.air

brigand

import React, { Component } from 'react';
import styled from 'styled-components';

const Figure = styled.figure`
  height: 0;
  margin: 0;
  background-color: #efefef;
  position: relative;
  padding-bottom: ${props => props.ratio}%;
`;

kmelve

.root {
  display: block;
  position: relative;
}

.lqip {
  image-rendering: pixelated;
  width: 100%;
  opacity: 1;
  transition: opacity 50ms 100ms ease-out;

iwa0

mlir-opt matmult.mlir -convert-linalg-to-loops -lower-affine -convert-scf-to-cf -convert-linalg-to-llvm -convert-memref-to-llvm -convert-func-to-llvm -reconcile-unrealized-casts > out.mlir
mlir-cpu-runner out.mlir -O3 -e main -entry-point-result=void --shared-libs=libmlir_runner_utils.dylib