Andrew Kane ankane

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

What I Wish I'd Known About Equity Before Joining A Unicorn

Disclaimer: This piece is written anonymously. The names of a few particular companies are mentioned, but as common examples only.

This is a short write-up on things that I wish I'd known and considered before joining a private company (aka startup, aka unicorn in some cases). I'm not trying to make the case that you should never join a private company, but the power imbalance between founder and employee is extreme, and that potential candidates would

Sending Bash and ZSH Commands to Syslog

Also posted here: http://18pct.com/sending-bash-and-zsh-commands-to-syslog/

Your bash/zsh history is great if its complete, but it doesn't capture commands across all users, sudo's, root commands etc. In particular with test environments, someone may perform a "one-off" procedure and then months later it needs to be repeated. It would be nice to be able to look up what the user did at the time, and searching through multiple, possibly truncated history files is a pain.

Tools like typescript are great if you're actively documenting, but not something you would use all the time in practice and capture more than just a history of your commands. There are third-party tools like rootsh and Snoopy that can accomplish this, but third-party tools can be overkill if all you want is a quick reference in a re

1. Gemfile

gem 'browserify-rails', '1.5.0' # until fix: https://github.com/browserify-rails/browserify-rails/issues/101
gem 'react-rails'

Browserify-rails allows to use browserify within assets pipeline. React-rails is here only to allow to use #react_component (and thus, prerendering).

Note that jquery-rails can be removed from Gemfile, the npm version of jquery and jquery-ujs will be used instead.

	This gist I write, because I couldn't find step by step instructions
	how to install and start postgresql locally (using conda within a conda environment - with the result
	that you can run it without sudo/admin rights on your machine!)
	and not globally in the operating system (which requires sudo/admin rights on that machine).

	I hope, this will help especially people new to postgresql (and those who don't have sudo/admin rights on a specific machine but want
	to run postgresql there)!

	####################################
	# create conda environment

	-- This creates an athena table that can parse ALB logs.
	-- Advantage of this over others are this works when the log ends with a trailing space
	-- plus it also breaks the http request into route and params for easier grouping

	CREATE EXTERNAL TABLE IF NOT EXISTS alb_logs (
	type string,
	timestamp string,
	elb string,
	client_ip string,
	client_port int,

	'''This script goes along the blog post
	"Building powerful image classification models using very little data"
	from blog.keras.io.
	It uses data that can be downloaded at:
	https://www.kaggle.com/c/dogs-vs-cats/data
	In our setup, we:
	- created a data/ folder
	- created train/ and validation/ subfolders inside data/
	- created cats/ and dogs/ subfolders inside train/ and validation/
	- put the cat pictures index 0-999 in data/train/cats

	create table mydata_real (id serial, date date, value text);
	create table mydata_real_y2015 (check (date >= '2015-01-01' and date < '2016-01-01')) inherits (mydata_real);
	create table mydata_real_y2016 (check (date >= '2016-01-01' and date < '2017-01-01')) inherits (mydata_real);

	create function mydata_nope() returns trigger language plpgsql
	as $f$ begin raise exception 'insert on wrong table'; return NULL; end; $f$;

	create trigger mydata_nope before insert on mydata_real execute procedure mydata_nope();

	create view mydata as select * from mydata_real;

	require 'securerandom'
	require 'concurrent/scheduled_task'
	require 'concurrent/executor/thread_pool_executor'
	require 'concurrent/utility/processor_counter'

	module ActiveJob
	module QueueAdapters
	# == Active Job Async adapter
	#
	# The Async adapter runs jobs with an in-process thread pool.

	import pandas as pd
	from random import random

	flow = (list(range(1,10,1)) + list(range(10,1,-1)))*100
	pdata = pd.DataFrame({"a":flow, "b":flow})
	pdata.b = pdata.b.shift(9)
	data = pdata.iloc[10:] * random() # some noise

	import numpy as np