MadFlow: automating Monte Carlo simulation on GPU for particle physics processes

We present MadFlow, a first general multi-purpose framework for Monte Carlo (MC) event simulation of particle physics processes designed to take full advantage hardware accelerators, in particular, graphics processing units (GPUs). The automation process generating all the required components MC generic and its deployment on accelerator is still big challenge nowadays. In order solve this challenge, we design workflow code library which provides user possibility simulate custom through MadGraph5_aMC@NLO plugin generation exporting specialized GPU-like format. exported includes analytic expressions matrix elements phase space. performed using VegasFlow PDFFlow libraries deploy automatically systems with different acceleration capabilities, such as multi-threading CPU, single-GPU multi-GPU setups. package also an asynchronous unweighted events procedure store results. Crucially, although only Leading Order automatized, ingredients necessary build complex simulators modern, extensible maintainable way. show results at leading-order multiple configurations.