Compiler-supported reduced precision and AoS-SoA transformations for heterogeneous hardware

TL;DR

This paper explores compiler-supported data layout transformations and reduced-precision computations for particle simulations on heterogeneous GPU platforms, aiming to optimize performance and data management.

Contribution

It introduces compiler annotations enabling efficient AoS-to-SoA transformations and reduced-precision data handling on GPUs, addressing data movement and precision trade-offs.

Findings

Nvidia G200 platforms achieve ~2.6x speedup with transformations.

AMD MI300A shows robust performance gains but less than Nvidia.

Compiler techniques are broadly applicable to Lagrangian codes.

Abstract

This study evaluates AoS-to-SoA transformations over reduced-precision data layouts for a particle simulation code on several GPU platforms: We hypothesize that SoA fits particularly well to SIMT, while AoS is the preferred storage format for many Lagrangian codes. Reduced-precision (below IEEE accuracy) is an established tool to address bandwidth constraints, although it remains unclear whether AoS and precision conversions should execute on a CPU or be deployed to a GPU if the compute kernel itself must run on an accelerator. On modern superchips where CPUs and GPUs share (logically) one data space, it is also unclear whether it is advantageous to stream data to the accelerator prior to the calculation, or whether we should let the accelerator transform data on demand, i.e.~work in-place logically. We therefore introduce compiler annotations to facilitate such conversions and to give the programmer the option to orchestrate the conversions in combination with GPU offloading. For some of our compute kernels of interest, Nvidia's G200 platforms yield a speedup of around 2.6 while AMD's MI300A exhibits more robust performance yet profits less. We assume that our compiler-based techniques are applicable to a wide variety of Lagrangian codes and beyond.