Abdallah al-swaiti

The Ultimate Guide to Building a High-Performance Linux Kernel with TKG

This guide provides a comprehensive, expert-level walkthrough for compiling a custom Linux kernel using the linux-tkg project. It is intended for enthusiasts and power users who want to move beyond generic distribution kernels to unlock the maximum performance, responsiveness, and low-latency potential of their hardware, particularly for desktop and gaming workloads.

The Philosophy: Why Build a Custom Kernel?

The kernel shipped by your Linux distribution is a masterpiece of compatibility, designed to run on everything from ancient laptops to multi-socket servers. This "one-size-fits-all" approach means it's not specifically optimized for your hardware or your use case.

By building a custom kernel, you are making a deliberate trade-off: sacrificing universal compatibility for specialized performance. You are creating a kernel that is explicitly engineered for your CPU, your hardware, and your workloads.

Arch Linux Installation Guide for Hyprland

Author: Abdallah Alswaiti
Last Updated: August 2025

Arch Linux BTRFS + LUKS2 Recovery Guide with Timeshift Snapshots

This guide covers how to recover a broken Arch Linux system using BTRFS snapshots (Timeshift) when your system won't boot or is corrupted. Based on a real recovery scenario with an encrypted NVME drive setup.

System Configuration Example

Drive: /dev/nvme0n1 (931.5GB)
Partitions:
- /dev/nvme0n1p1 - 512MB EFI boot partition (FAT32)
/dev/nvme0n1p2 - 931GB encrypted partition (LUKS2)

Improve Fonts

Newest

Make your Arch fonts beautiful easily! This is what I do when I install Arch Linux to improve the fonts.

You may consider the following settings to improve your fonts for system-wide usage without installing a patched font library packages (eg. Infinality):

Install some fonts, for example:
sudo pacman -S ttf-dejavu ttf-liberation noto-fonts

	# --- الخطوة 1: تعريف الأدوات الوهمية (Dummy Tools) ---
	# في تطبيق حقيقي، قد تتصل هذه الدوال بواجهات برمجية (APIs) حقيقية

	def get_weather(text_input):
	"""أداة وهمية لمعرفة الطقس في مدينة ما"""
	# منطق بسيط جدًا لاستخراج اسم المدينة (للتوضيح فقط)
	if "in" in text_input:
	city = text_input.split("in")[-1].strip()
	return f"The weather in {city} is sunny and pleasant."
	return "Please specify a city, for example: 'weather in Riyadh'."

	import os
	import torch
	import gguf
	import numpy as np
	import argparse
	from tqdm import tqdm
	from safetensors.torch import load_file, save_file

	CHUNK_SIZE = 100 # Number of tensors to process in each chunk
	QUANTIZATION_THRESHOLD = 1024

	# https://huggingface.co/PixArt-alpha/PixArt-Sigma-XL-2-1024-MS/tree/main/transformer
	# https://huggingface.co/blog/quanto-diffusers

	#python -m venv quanta or conda whatever
	#source ./quanta/bin/activate >>> default (bash shell)


	# pip install --force git+https://github.com/huggingface/diffusers.git
	# pip install torch optimum.quanto
	#run first part to create model

	<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" viewBox="0 0 750 500" preserveAspectRatio="xMidYMax slice">
	<defs>
	<!-- Scene 1 Gradient -->
	<linearGradient id="grad1" x1="-154.32" y1="263.27" x2="-154.32" y2="374.3"
	gradientTransform="matrix(-1, 0, 0, 1.36, 231.36, -100.14)" gradientUnits="userSpaceOnUse">
	<stop offset="0.07" stop-color="#9c536b" />
	<stop offset="0.98" stop-color="#d98981" />
	</linearGradient>
	<radialGradient id="bg_grad" cx="375" cy="-35" r="318.69" gradientUnits="userSpaceOnUse">
	<stop offset="0.1" stop-color="#F5C54E" id="sun" />

	<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" viewBox="0 0 750 500" preserveAspectRatio="xMidYMax slice">
	<defs>
	<!-- Scene 1 Gradient -->
	<linearGradient id="grad1" x1="-154.32" y1="263.27" x2="-154.32" y2="374.3"
	gradientTransform="matrix(-1, 0, 0, 1.36, 231.36, -100.14)" gradientUnits="userSpaceOnUse">
	<stop offset="0.07" stop-color="#9c536b" />
	<stop offset="0.98" stop-color="#d98981" />
	</linearGradient>
	<radialGradient id="bg_grad" cx="375" cy="-35" r="318.69" gradientUnits="userSpaceOnUse">
	<stop offset="0.1" stop-color="#F5C54E" id="sun" />

	<div class="container">
	<p>/><span class="typed-text"></span><span class="cursor"> </span></p>
	</div>