Matthew Bonanni

A high-throughput and memory-efficient inference and serving engine for LLMs

Unstructured, GPU accelerated, high-order finite volume CFD solver

Interactive visualizer for transformer self-attention variants (MHA, GQA, MQA, MLA) with tensor shapes, FLOPs/memory cost analysis, and per-GPU roofline estimates

Pareto multiobjective optimization software

Calculates projectile motion with linear and quadratic models of air resistance

LLM inference project

Computes and tracks the accuracy of a mechanical watch

FlashMLA: Efficient MLA kernels

Fast and memory-efficient exact attention

No description.

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Common recipes to run vLLM

A simple GPU reservation tool for single host shared development systems

This repo hosts code for vLLM CI & Performance Benchmark infrastructure.

Tally - Personal Finance App

Spitfire is a Python/C++ library for constructing tabulated chemistry models and solving differential equations.

simple library for numerical continuation based on NumPy

Final project for AA228

CS230 Final Project: Shock Parameterization for the Compressible Euler Equations using Deep Learning

A fire growth model

Percolation model of wildfire spread implemented in Matlab

This software models and optimizes the accessible volume ("gamut") of arbitrary kinematic chains when constrained by a glovebox.

MATLAB analysis software for ME 4550 final design project

Chat client developed in satisfaction of the requirements of CS 2500

Solver for soma cube-type puzzles