Search NASA⌕ Search

Engineering topics

Tramm, John

Publications and source records attributed to Tramm, John.

MONTE CARLO CROSS SECTION LOOKUP KERNEL FOR THE CEREBRAS WSE-2 IN CSL

This is a small kernel that was used to collect data for an upcoming paper. We would like to have the code be open source so that the reviewers (and then readers) of the paper can see the whole code, and can reproduce/verify our results. This is not a fully featured application, it cannot produce any useful simulation results, it just executes a small abstracted kernel using synthetic data. The purpose of the kernel is to understand the basic performance characteristics of an HPC kernel on novel AI accelerator architectures. The main kernel is written in the CSL coding language for use with the Cerebras WSE-2 AI accelerator. The kernel represented is the Monte Carlo cross section lookup kernel, which is a small kernel used by the Monte Carlo neutral particle transport algorithm. There is also a baseline kernel written in CUDA that we will include in the repository to form a basis for comparing the WSE-2 to GPU.

Tramm, John↗

Just-in-Time Compilation and Link-Time Optimization for OpenMP Target Offloading

Following the mass adoption of external accelerators for high performance computing, the overall performance of many applications has become increasingly dependent on relatively small accelerated kernels. As static analysis is fundamentally limited by dynamic values and external definitions, standard ahead-of-time compilation is not always sufficient to achieve the best performance. Furthermore, many users looking to port an existing application to run on an external accelerator will not want to fundamentally restructure their programs. These and other problems can be addressed through both link-time optimization (LTO) and just-in-time (JIT) compilation, but until now had sparse and inconsistent support from the compiler. In this work, we present a new compilation method that enables device-side LTO as well as a transparent JIT compilation tool-chain for OpenMP target offloading. Our contributions include an entirely new device linking and embedding scheme to enable LTO as well as a novel JIT engine to efficiently optimize OpenMP offloading regions at run-time. We also introduce a persistent caching system to improve end-to-end runtime using the JIT engine and minimize kernel launching overheads. We measure the performance of our LTO and JIT implementation via several real-world scientific applications. With our optimizations we observe significant improvements through LTO on large applications as well as significant end-to-end execution time improvement using JIT.

Tian, Shilei↗