% ============================================================================
% UQO Working Paper No. 40 — RAID: A Metal-Accelerated C++ Framework for Reproducible
% High-Frequency Financial Econometrics
% ============================================================================
\documentclass[12pt,letterpaper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[english]{babel}
\usepackage[letterpaper,top=1in,bottom=1in,left=1in,right=1in]{geometry}
\usepackage{mathptmx}
\usepackage{microtype}
\usepackage{setspace}
\onehalfspacing
\usepackage{amsmath,amssymb,amsthm}
\usepackage{booktabs}
\usepackage{tabularx}
\usepackage{multirow}
\usepackage{threeparttable}
\usepackage{longtable}
\usepackage{siunitx}
\sisetup{input-symbols = {()}, table-format = -1.3, parse-numbers = false}
\usepackage{graphicx}
\usepackage{subcaption}
\usepackage{float}
\usepackage[dvipsnames]{xcolor}
\usepackage[colorlinks=true,linkcolor=NavyBlue,citecolor=NavyBlue,urlcolor=NavyBlue,breaklinks=true]{hyperref}
\usepackage[authoryear,round,semicolon]{natbib}
\bibliographystyle{aer}
\usepackage[font=small,labelfont=bf,labelsep=period,justification=justified,singlelinecheck=false]{caption}
\usepackage{fancyhdr}
\pagestyle{fancy}
\fancyhf{}
\renewcommand{\headrulewidth}{0pt}
\fancyfoot[C]{\thepage}
\usepackage{titlesec}
\titleformat{\section}{\large\bfseries}{\thesection.}{0.5em}{}
\titleformat{\subsection}{\normalsize\bfseries}{\thesubsection.}{0.5em}{}
\titleformat{\subsubsection}{\normalsize\itshape}{\thesubsubsection.}{0.5em}{}
\usepackage[toc,page]{appendix}
\usepackage{enumitem}
\setlist{nosep,leftmargin=*}
\usepackage{listings}
\lstset{basicstyle=\ttfamily\footnotesize,breaklines=true,columns=fullflexible,frame=single,framerule=0.2pt}
\usepackage{url}
\newtheorem{proposition}{Proposition}
\newcommand{\RV}{\mathrm{RV}}
\newcommand{\BPV}{\mathrm{BPV}}
\newcommand{\RQ}{\mathrm{RQ}}
\newcommand{\RSV}{\mathrm{RSV}}
\newcommand{\raid}{\textsc{raid}}
\newcommand{\E}{\mathbb{E}}
\newcommand{\I}{\mathbb{1}}

% ============================================================================
% METADATA
% ============================================================================
\newcommand{\WPnumber}{40}
\newcommand{\WPtitle}{RAID: A Metal-Accelerated C++ Framework for Reproducible High-Frequency Financial Econometrics}
\newcommand{\WPsubtitle}{Batched, Fused and Streaming Estimation on Apple Silicon without Sacrificing Inference}
\newcommand{\WPdate}{November 2026}
\newcommand{\WPversion}{1.0}
\newcommand{\WPabstract}{%
  We ask whether heterogeneous high-frequency econometric workloads---realized measures, rolling and batched regressions, HAR forecasting models, GARCH-family quasi-maximum likelihood, block bootstrap inference---can be reorganised into batched, fused, streaming computation on Apple Silicon GPUs without sacrificing estimator accuracy, statistical inference, or reproducibility. Apple GPUs expose no 64-bit floating point. \raid{} answers with a C++20 library whose FP64 CPU backend is the reference, whose Metal backend accumulates every statistic in compensated ``double-float'' arithmetic and combines chunked partial sums in FP64, and whose estimators report backend, precision, convergence and provenance in every result. Tier~A methods reproduce \texttt{highfrequency}, \texttt{rugarch}, \texttt{statsmodels} and \texttt{arch} to documented tolerances; Monte Carlo experiments confirm bias, RMSE and coverage; GPU and CPU results agree to $10^{-6}$ relative and are bitwise reproducible across runs and thread counts. Using one year of 1-minute bars for a seeded universe of 1,000 US equities served by a private FirstRate Data lakehouse, we measure break-even points on an M5~Max laptop and across eleven Apple Silicon generations (M2 to M6). The GPU wins on fused realized measures beyond roughly $4\times10^4$ elements and on batched GARCH likelihood evaluation when many parameter vectors are evaluated per series; the CPU wins on rolling sums, on single-vector GARCH recursions and on small panels. Conversion and buffer traffic, not kernel time, dominate GPU wall-clock: unified memory is not free computation. We publish the negative results, the tolerances, and a reproducible benchmark protocol.%
}
\newcommand{\WPkeywords}{high-frequency econometrics, realized volatility, HAR, GARCH, GPU computing, Apple Metal, reproducibility, numerical precision}
\newcommand{\WPjel}{C58, C63, C87, G17}
\newcommand{\WPauthor}{Simon-Pierre Boucher}
\newcommand{\WPaffiliation}{Département des sciences administratives\\Université du Québec en Outaouais}
\newcommand{\WPemail}{simon-pierre.boucher@uqo.ca}
\newcommand{\WPaddress}{Gatineau -- Pavillon Alexandre-Taché\\283, boulevard Alexandre-Taché\\Gatineau, Québec, Canada J9A 1L8}

\begin{document}
\input{sections/titlepage}
\input{sections/introduction}
\input{sections/literature}
\input{sections/architecture}
\input{sections/estimators}
\input{sections/metal}
\input{sections/correctness}
\input{sections/setup}
\input{sections/microbenchmarks}
\input{sections/endtoend}
\input{sections/applications}
\input{sections/threats}
\input{sections/reproducibility}
\input{sections/conclusion}
\newpage
\bibliography{references}
\input{appendices/appendix}
\end{document}