Pointers Inside Lambda Closure Objects in OpenMP Target Offload Regions

Abstract

With the diversification of HPC architectures beyond traditional CPU-based clusters, a number of new frameworks for performance portability across architectures have arisen. One way of implementing such frameworks is to use C++ templates and lambda expressions to design loop-like functions. However, lower level programming APIs that these implementations must use are often designed with C in mind and do not specify how they interact with C++ features such as lambda expressions. This paper discusses a change to the behavior of the OpenMP specification with respect to lambda expressions such that when functions generated by lambda expressions are called inside GPU regions, any pointers used in the lambda expression correctly refer to device pointers. This change has been implemented in a branch of the Clang C++ compiler and demonstrated with two representative codes. This change has also been accepted into the draft OpenMP ® specification for inclusion in OpenMP 5. Our results show that the implicit mapping of lambda expressions always exhibits identical performance to an explicit mapping but without breaking the abstraction provided by the high level frameworks.

Developing performance portable plasma edge simulations: A survey

“…Support for target offload can be found in commercial compilers from Intel, IBM, AMD and Cray, with a variety of target architectures. Support also exists in the Clang/LLVM [56] and GCC open-source compilers. 2 While the explicit device control provided by the CUDA, ROCm/HIP, and OpenCL programming models may be more powerful than directivebased approaches, it may also significantly increase developer effort.…”