Nico Weber nico

🖥️ See Demo

3.5 fps, Paperwhite 3
@adtac_

step 1: jailbreak your Kindle

mobileread.com is your best resource here, follow the instructions from the LanguageBreak thread

I didn't really follow the LanguageBreak instructions because I didn't care about most of the features + I was curious to do it myself, but the LanguageBreak github repo was invaluable for debugging

As per gfx-rs/gfx#3716 (comment) :

macOS has a system shader cache at $(getconf DARWIN_USER_CACHE_DIR)/com.apple.metal

On my MacBookPro that is under /var/folders/52/l9z1nqld5yg99tb_s3q6nyhh0000gn/C

System shaders: /var/folders/52/l9z1nqld5yg99tb_s3q6nyhh0000gn/C/com.apple.metal
Blender shaders: /var/folders/52/l9z1nqld5yg99tb_s3q6nyhh0000gn/C/org.blenderfoundation

Delete all the folders in there to clear the cache.

Anti-hype LLM reading list

Goals: Add links that are reasonable and good explanations of how stuff works. No hype and no vendor content if possible. Practical first-hand accounts of models in prod eagerly sought.

Foundational Concepts

Pre-Transformer Models

Helping your 'old' PC build faster with your mobile device

It all happened when I decided to run Geekbench 5 on my phone: surprisingly the single-core performance matched my 'old'¹ Pentium T3200 and surpassed it in multicore. Since I've been having fun with distcc for the last few days, I asked myself: 'Can my phone really help my old laptop build faster? nah, hard to believe... but let's try'.

Without further ado: YES. Not only can my phone be faster, but it can significantly help in the build process, I believe the results below speak for themselves:

Building Git (#30cc8d0) on a Pentium T3200, 8m30s

Building Git (#30cc8d0) on a Pentium T3200 (2x 2.0 GHz)+ Snapdragon 636 (4x1.8 + 4x1.6 GHz), 2m9s

	#!/usr/bin/env python3

	from collections import namedtuple
	from pathlib import Path
	import base64
	import httplib2
	import json
	import os
	import subprocess

	// Sometimes you have a large file on a small disk and would like to "transform"
	// it in some way: for example, by decompressing it. However, you might not have
	// enough space on disk to keep both the the compressed file and the
	// decompressed results. If the process can be done in a streaming fashion, it
	// would be nice if the file could be "drained"; that is, the file would be
	// sequentially deleted as it is consumed. At the start you'd have 100% of the
	// original file, somewhere in the middle you'd have about half of the original
	// file and half of your output, and by the end the original file will be gone
	// and you'll be left with just the results. If you do it this way, you might
	// be able to do the entire operation without extra space!

	# Classic Fork-join parallelism. bash_parallel's can be nested arbitrarily
	# Silent by default; set BASH_PARALLEL_VERBOSE=1 for verbose output on stderr
	function bash_parallel
	{
	function bash_parallel_echo
	{
	if [[ $BASH_PARALLEL_VERBOSE == 1 ]] ; then
	echo "$@" 1>&2
	fi
	}

	# Clone llama.cpp
	git clone https://github.com/ggerganov/llama.cpp.git
	cd llama.cpp

	# Build it
	make clean
	LLAMA_METAL=1 make

	# Download model
	export MODEL=llama-2-13b-chat.ggmlv3.q4_0.bin

	" Pick a different highlighting group by adding this in your vimrc:
	"
	" highlight link fuzzyMatch <group>
	"
	" The "default" only makes the link if it's not already set.
	"
	highlight default link fuzzyMatch Search

	" The components of the command definition:
	"