Performance Portability Strategies for Grid C++ Expression Templates

Abstract

Abstract. One of the key requirements for the Lattice QCD Application Development as part of the US Exascale Computing Project is performance portability across multiple architectures. Using the Grid C++ expression template as a starting point, we report on the progress made with regards to the Grid GPU offloading strategies. We present both the successes and issues encountered in using CUDA, OpenACC and Just-In-Time compilation. Experimentation and performance on GPUs with a SU(3)×SU(3) streaming test will be reported. We will also report on the challenges of using current OpenMP 4.x for GPU offloading in the same code.

Lyncs-API: a Python API for Lattice QCD applications

“…For example, the QUDA library [1] is implementing various backends for extending its employment on GPUs made by other manufacturers than NVIDIA, such as AMD or Intel [2]. And the Grid library, whose development focused on achieving portability on CPU architectures [3], has been also implementing GPU support [4]. There are a handful of more examples where the lattice QCD community is either extending or developing new software for adapting to the diversity of computing architectures available nowadays.…”

Lattice QCD on new chips: a community summary

“…M. Lin presented successes and challenges encountered in porting the Grid C++ expression template to GPU-based systems and exploring extensively different approaches to integrate CUDA, Ope-nACC and Just-In-Time compilation [22].…”

SVE-Enabling Lattice QCD Codes

“…This intrinsic function returns the SVE vector register length (in double). In the loop body, we use the intrinsic function svld1() to load slices of the arrays x and y without decomposing the array elements (lines [8][9]. Computation proceeds with multiplyadd of complex numbers using two calls to svcmla() (the intrinsic function for the FCMLA instruction introduced in Section III-D) (lines 10-11).…”

“…A more general approach targeting various x86 SIMD ISAs, but in particular AVX-512, is Grid [4]. Meanwhile exploratory studies have been performed to extend the portability of Grid to other types of architectures, including GPU-accelerated ones [8]. A much earlier effort targeting architectures comprising NVIDIA GPUs supporting CUDA resulted in the QUDA library [3], which has meanwhile been used for several generations of GPU-accelerated supercomputers.…”