rygorous’s gists

rygorous / more_serious_fft.cpp

Created December 10, 2025 06:11

Core for a 2x unrolled radix-2 (not really radix-4) FFT kernel

	// The FFT alg used here was designed to be very FMA-friendly, but because we can't assume FMAs are present on
	// all target HW and want consistent results everywhere, we're using FMA-less algorithms for this application.

	// Notation throughout this file:
	//
	// Let z = a + bi. Then conj(z) = a - bi.
	//
	// We can swap the real and imaginary parts of z to yield s(z) = b + ai ("swap").
	// Now because
	//

rygorous / gist:9aac91598af0a94ab4693210291e1f94

Created December 10, 2025 05:37

Reference r2 FFT code

	static size_t const kMaxN = 2048; // This is the largest straight FFT we support

	struct complexf
	{
	float re;
	float im;

	complexf() {}
	complexf(float r) : re(r), im(0.0f) {}
	complexf(float r, float i) : re(r), im(i) {}

rygorous / morton3d_fun.cpp

Created July 4, 2025 05:49

Vectorized 3D Morton encoding for no reason

	#include <immintrin.h>
	#include <stdio.h>
	#include <stdint.h>
	#include <assert.h>

	typedef uint32_t uint32;
	typedef __m128i Vec128_U32;

	// "Insert" two 0 bits after each of the 11 low bits of x
	static uint32 Part1By2(uint32 x)

rygorous / gist:6f96cc21292cc704f53ef77e5b4be519

Created May 22, 2025 02:25

Oodle Texture BC7RD "preserve extremes" constraint validation

	//===================================================================
	// constraint validation for preserve extremes mode

	// Cold part of are_endpoitns_permitted: the actual constraint validation
	static RADNOINLINE bool are_endpoints_permitted_cold(const BC7BlockState & st, const bc7rd_blockinfo& info)
	{
	const BC7Flags flags = info.flags;

	if (st.mode <= 3)
	{

rygorous / gist:52cc2a23a73813d645581046dced27fd

Created May 22, 2025 00:15

Oodle Texture "preserve extremes" rules

	{
	BC7Flags flags = in_flags;

	// Preserve extremes mode.
	//
	// This mode preserves values of 0 and 255 in the alpha channel exactly. In general,
	// we can do this in any one channel, but restricting this to alpha makes the interface
	// simpler, is consistent with what we do for BC3, and doesn't seem like a signfiicant
	// limitation for the user.
	//

rygorous / gist:94fd3c33afe4e650ec5bda386cbc16bb

Created May 8, 2025 01:16

	uint EvenBitMask = 0x55555555u;
	uint HighIndexBit = PackedIndices >> 1;

	float NumWeight1 = float(countbits(EvenBitMask & ~HighIndexBit & PackedIndices));
	float NumWeight2 = float(countbits(EvenBitMask & HighIndexBit & ~PackedIndices));
	float NumWeight3 = float(countbits(EvenBitMask & HighIndexBit & PackedIndices));

	// with NV LOP3: 1 shift, 3 LOP3, 3 pop count, 3 int->float = 10 insns total for 16 pixels
	// without: 1 shift, 2 NOT, 5 AND, 3 pop count, 3 int->float = 14 insns total for 16 pixels

rygorous / rr_dds.h

Created March 7, 2025 01:45

rr_dds loader/writer

	//===================================================
	// Oodle2 DDS Tool
	// (C) Copyright 1994-2022 Epic Games Tools LLC
	//===================================================

	#ifndef RR_DDS_INCLUDED
	#define RR_DDS_INCLUDED

	#include <stdint.h>
	#include <stddef.h>

rygorous / main.rs

Created March 2, 2025 04:42

Base64 fixed point test

	use bit_set::BitSet;
	use std::mem;

	// Vanilla RFC 4648
	const ALPHABET: &str = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
	// URL-safe RFC 4648
	//const ALPHABET: &str = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_";
	const COUNT: usize = 1usize << 24; // 3 bytes worth suffices for this test

	fn lookup(index: u32) -> u32 {

rygorous / bc7rd_chart_example.html

Created January 2, 2025 09:46

Example run with a BC7 RD plot diff

	<!DOCTYPE html>
	<html><head>
	<meta http-equiv="content-type" content="text/html; charset=UTF-8">
	<style>
	body { background: #fff; margin: 10px; font: 12pt Calibri,Arial,Helvetica,sans-serif; }
	h1 { font-size: 18pt; margin: .5em 0 .5em 0; }
	h2 { font-size: 16pt; margin: .5em 0 .5em 0; }
	canvas { margin: auto; }
	table { border-spacing: 0; }
	table.rdresult tbody tr:nth-child(even) { background: #eee; }

rygorous / rdchart_template_prefix.html

Created January 2, 2025 09:40

RD chart template for Oodle Texture eval

	<!DOCTYPE html>
	<html>
	<head>
	<style>
	body { background: #fff; margin: 10px; font: 12pt Calibri,Arial,Helvetica,sans-serif; }
	h1 { font-size: 18pt; margin: .5em 0 .5em 0; }
	h2 { font-size: 16pt; margin: .5em 0 .5em 0; }
	canvas { margin: auto; }
	table { border-spacing: 0; }
	table.rdresult tbody tr:nth-child(even) { background: #eee; }

Fabian 'ryg' Giesen rygorous