DSATS : Research & Project

Project 1 : The GAMES Exascale Computing Project

Heterogeneous catalysis and the design of new catalysts is a grand challenge problem in computational chemistry that will require the capabilities of exascale computing. The GAMESS project is extending methods and algorithms based on chemical fragmentation methods and coupling these with high-fidelity quantum chemistry (QC) simulations to solve this problem.

As part of the Exascale Computing Project, GAMES is currently being refactored to take advantage of modern computer hardware and software, and the capabilities of the C++ libcchem code that is co-developed with GAMES are being greatly expanded.

• Our group is at the forefront of the GAMES ECP initiative, which is a concerted effort of the Oak Ridge and Argonne Leadership Computing Facilities, the Ames National Lab, and various other US collaborators and vendors.
• Our development in this project involves leveraging novel hardware architectures and programming model with the ultimate goal of devising software that can be executed on the exascale machines Frontier and Aurora to push the edge of what is currently achievable in chemical modelling.

Project 2 : Architecture and Structure Aware Linear Algebra

• Linear algebra (LA) operations are fundamental to a large number of computational science algorithms.
• LA algorithms is complicated by the increasing architectural heterogeneity of the high-performance computing (HPC) platforms.
• This project aims to build an Architecture and Data-Structure Aware Linear Algebra (ADSALA) software package that will use machine learning to learn the hardware/data-structure/package/algorithm relationships when compiled on a specific hardware architecture for a spectrum of LA packages.

The pursuit of optimal LA algorithms is significantly complicated by the increasing architectural heterogeneity of the high-performance computing (HPC) platforms, with a variable mix of general-purpose processors (CPUs) and accelerators (GPUs, DSPs, FPGAs, etc.), and complex associated memory hierarchies and file systems.

Linear algebra (LA) operations are fundamental to a large number of computational science algorithms. The applications span the entire scientific board, with machine learning (ML) algorithms being among the most reliant on LA operations; they provide the mathematics that underpins much of what we do. Historically, this fact has driven the development of a plethora of libraries providing high-performance implementations of LA algorithms: BLAS, OpenBLAS, cuBLAS, CLBLAS, LAPACK, ARPACK, ATLAS, cuSOLVER, MAGMA and many more. For a given LA operation, the choice can be bewildering for the programmer, especially given that within the same library there may be several algorithms yielding different performance depending, for example, on the specific structure of the matrices involved.

This project aims to build an Architecture and Data-Structure Aware Linear Algebra (ADSALA) software package that will use machine learning to learn the hardware/data-structure/package/algorithm relationships when compiled on a specific hardware architecture for a spectrum of LA packages. At runtime, after analysing the structural features of the data structures involved, ADSALA will choose the most appropriate package/algorithm for a given LA operation, and assign the computation to the best combination of hardware resources (CPU, GPU), seeking to minimize execution time – all subject to specific user-defined constraints.