hellerbarde

Latency numbers every programmer should know

L1 cache reference ......................... 0.5 ns
Branch mispredict ............................ 5 ns
L2 cache reference ........................... 7 ns
Mutex lock/unlock ........................... 25 ns
Main memory reference ...................... 100 ns             
Compress 1K bytes with Zippy ............. 3,000 ns  =   3 µs
Send 2K bytes over 1 Gbps network ....... 20,000 ns  =  20 µs
SSD random read ........................ 150,000 ns  = 150 µs

Read 1 MB sequentially from memory ..... 250,000 ns = 250 µs

fzdwx

import java.util.concurrent.locks.LockSupport;

public class test {

    public static void main(String[] args) {
        final Thread t6 = new Thread(new T(6, null));
        final Thread t5 = new Thread(new T(5, t6));
        final Thread t4 = new Thread(new T(4, t5));
        final Thread t3 = new Thread(new T(3, t4));
        final Thread t2 = new Thread(new T(2, t3));

fzdwx

package main

import (
	"fmt"
	"os"
)

func main() {

	c1 := make(chan int, 1)

lxzan

package bench

import (
	"github.com/bytedance/gopkg/util/gopool"
	"github.com/lesismal/nbio/taskpool"
	ants "github.com/panjf2000/ants/v2"
	"github.com/redis/go-redis/v9"
	"sync"
	"testing"
)

ninehills

karpathy

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.