Raimo33

#!/usr/bin/env bash
set -e

N=${1:-1}
FILTER=${2:-}
TIME=${3:-}

CURRENT_COMMIT=$(git rev-parse HEAD)
COMMITS=$(git rev-list --max-count="$N" --reverse HEAD)

Raimo33

#!/usr/bin/env bash
set -e

N=${1:-1}

CURRENT_COMMIT=$(git rev-parse HEAD)
COMMITS=$(git rev-list --max-count="$N" --reverse HEAD)

trap 'git checkout -q "$CURRENT_COMMIT"' EXIT SIGINT SIGTERM

Raimo33

FROM debian:trixie-slim AS builder

RUN apt-get update && apt-get install -y \
        build-essential \
        pkg-config \
        cmake \
        ninja-build \
        libboost-system-dev \
        libboost-filesystem-dev \
        libboost-program-options-dev \

Raimo33

#!/usr/bin/env bash
set -u

out_file="benchmarks_throwing_exception.txt"
log="$(mktemp)"

mapfile -t benches < <(./build/bin/bench_bitcoin --list)

rm -f "$out_file"

Raimo33

template <typename T, typename Comparator>
HOT ssize_t forward_lower_bound(std::span<const T> data, const T elem, const Comparator &comp) noexcept
{
  static_assert(std::is_integral<T>::value, "T must be an integral type");
  static_assert(std::hardware_constructive_interference_size == 64, "Cache line size must be 64 bytes");

  static constexpr uint8_t chunk_size = sizeof(__m512i) / sizeof(T);
  const T *begin = data.data();
  const size_t size = data.size();
  size_t remaining = size;

Raimo33

/*
description:
  takes pointers to the head and tail of the singly-linked list.
  sorts the list in ascending order by ASCII value.

implementation:
  use of mergesort because linked lists inherently don't provide random access. could be bypassed by constructing an array of pointers but that's a poor design choice O(n)...
  merge sort is the fastest known sorting algo for linked lists, even though finding the midpoint is O(n/2) = O(n).
  use of the slow-fast pointer technique to find the midpoint of the list. same performance as the doubly-linked list -> <- approach.
  implemented by changing pointers, even though a char is 1 byte and easier to copy, for better mantainability, reusability, and standard compliance.

Raimo33

//TODO find a way to make them const, forcing prevention of thread safety issues, constexpr??
#ifdef __AVX512F__
  static __m512i _512_vec_A_minus_1;
  static __m512i _512_vec_case_range;
  static __m512i _512_add_mask;
#endif

#ifdef __AVX2__
  static __m256i _256_vec_A_minus_1;
  static __m256i _256_vec_case_range;

Raimo33

C Speed Optimization Checklist

This is a list of general-purpose optimizations for C programs, from the most impactful to the tiniest low-level micro-optimizations to squeeze out every last bit of performance. It is meant to be read top-down as a checklist, with each item being a potential optimization to consider. Everything is in order of speed gain.

Algorithm && Data Structures

Choose the best algorithm and data structure for the problem at hand by evaluating:

1. time complexity

Raimo33

#!/usr/bin/env python3

import os
import subprocess
import datetime
import re

CREATOR_NAME = "Claudio Raimondi"
CREATOR_EMAIL = "[email protected]"
HEADER_WIDTH = 80

Raimo33

# **************************************************************************** #
#                                                                              #
#                                                         :::      ::::::::    #
#    Makefile                                           :+:      :+:    :+:    #
#                                                     +:+ +:+         +:+      #
#    By: craimond <[email protected]>            +#+  +:+       +#+         #
#                                                 +#+#+#+#+#+   +#+            #
#    Created: 2024/08/06 01:09:08 by craimond          #+#    #+#              #
#    Updated: 2024/09/15 14:24:42 by craimond         ###   ########.fr        #
#                                                                              #