Discover gists

ASUS ROG Zephyrus G14 2022 GA402RJ pci passthrough

This is a record for my own passthrough setup, I can finally use a single laptop for windows gaming and linux programming at the same time.

The device

My G14 is GA402RJ(6800HS + 6700s) with MT7922 WiFi/BT card.

At this time, BIOS version is 309.

Docker: Wiki.js and PostgreSQL

Docker Swarm

https://docs.docker.com/engine/swarm/swarm-tutorial/

swarm
-----
    | - manager (alpha) [8GB HDD1/, 32GB HDD2/data, 4 Cores, 6GB RAM]

Introduction

In CS2, when a player is sliding/surfing against some geometry, it's possible that they lose all their momentum/velocity. This is usually called wallbug/rampbug. While the name was inherited from Source 1 games (such as CS:GO, CS:S, TF2,...), rampbugs in CS2 behave much differently from its predecessor. In contrast to source 1 games, it's also possible have the velocity redirected to another direction instead of losing all momentum.

This bug doesn't seem to be common in all source 2 games (HL:A for instance, does not have this bug), and was orignially not present in the very early versions of CS2 Limited Test. Rampbugs become more and more frequent over time, with the Call to Arms update effectively doubling the frequency of these bugs, which is a significant problem for custom gamemodes heavily depending on geometry collision (eg. surf).

Keep in mind that while the player collision hitbox is a box, the images shown below will represent the player as a dot instead for simplicity.

Observati

Moved to git repository: https://github.com/denji/nginx-tuning

NGINX Tuning For Best Performance

For this configuration you can use web server you like, i decided, because i work mostly with it to use nginx.

Generally, properly configured nginx can handle up to 400K to 500K requests per second (clustered), most what i saw is 50K to 80K (non-clustered) requests per second and 30% CPU load, course, this was 2 x Intel Xeon with HyperThreading enabled, but it can work without problem on slower machines.

You must understand that this config is used in testing environment and not in production so you will need to find a way to implement most of those features best possible for your servers.

`vs` Questions

http://stackoverflow.com/questions/804115 (rebase vs merge).
https://www.atlassian.com/git/tutorials/merging-vs-rebasing (rebase vs merge)
https://www.atlassian.com/git/tutorials/undoing-changes/ (reset vs checkout vs revert)
http://stackoverflow.com/questions/2221658 (HEAD^ vs HEAD~) (See git rev-parse)
http://stackoverflow.com/questions/292357 (pull vs fetch)
http://stackoverflow.com/questions/39651 (stash vs branch)
http://stackoverflow.com/questions/8358035 (reset vs checkout vs revert)

News

Important: Dev Mode app patch and alternatives (2024-04-14)

Since approximately 2024-04-02, the latest LG Dev Mode app (2.1.2 in the Content Store) copies and resets the permissions of jail_app.conf/jail_app.conf.sig on every boot. Therefore, jailpatch.sh etc. will no longer work.

If you have webOS 5+ and old enough firmware, WTA (which does not require Dev Mode) will still work.

If you have webOS 4.x, you can try CVE-2023-6319. It is unpatched on the latest (final?) firmware for webOS 4.0 (2018) models.

	#!/bin/bash

	CONTAINER=$1

	RUNNING=$(docker inspect --format="{{ .State.Running }}" $CONTAINER 2> /dev/null)

	if [ $? -eq 1 ]; then
	echo "'$CONTAINER' does not exist."
	else
	/usr/bin/docker rm --force $CONTAINER

	#!/usr/bin/python
	import os.path
	import re
	import os
	import sys
	import copy
	import eyed3
	from eyed3.id3.tag import Tag

	helptext = """

	# Apache License 2.0
	# 使用法は gist のコメントを見てください
	import argparse
	from typing import List, Optional, Union, Iterator

	from llama_cpp.llama_chat_format import _convert_completion_to_chat, register_chat_completion_handler
	import llama_cpp.llama_types as llama_types
	from llama_cpp.llama import LogitsProcessorList, LlamaGrammar
	from llama_cpp import Llama, llama_chat_format
	import gradio as gr

	# Thanks https://discuss.pytorch.org/t/rmse-loss-function/16540

	class RMSELoss(torch.nn.Module):
	def __init__(self):
	super(RMSELoss,self).__init__()

	def forward(self,x,y):
	criterion = nn.MSELoss()
	loss = torch.sqrt(criterion(x, y))
	return loss