Refine
Document Type
- In Proceedings (9) (remove)
Language
- English (9)
Has Fulltext
- no (9)
Is part of the Bibliography
- no (9)
Institute
For writing a new scientific application, portability across existing and future hardware should be the major design goal, as there is a multitude of different compute devices, and programme codes typically outlive systems by far. Unlike other programming models that address parallelism or heterogeneity, OpenCL does provide practical portability across a wide range of HPC-relevant architectures. Other than that, it has a range of further advantages like being a library-only implementation, and using runtime kernel-compilation. We present experiences with utilising OpenCL alongside C++, MPI, and CMake in two real-world scientific codes. Our targets are a Cray XC40 supercomputer with multi- and many-core (Xeon Phi) CPUs, as well as multiple smaller systems with Nvidia and AMD GPUs. We shed light on practical issues arising in such a scenario, like the interaction between OpenCL and MPI, discuss solutions, and point out current limitations of OpenCL in the domain of scientific HPC from an application developer's and user's point of view.
The NEC SX-Aurora TSUBASA is a new generation of vector processing architectures that combines a standard Intel Xeon host with the newly developed NEC Vector Engine co-processor cards. One way to use these co-processors is offloading suitable parts of the program from the host to the Vector Engines. Currently, the only vendor-provided offloading solutions are the low-level Vector Engine Offloading (VEO) library, and a builtin reverse-offloading mechanism named VHcall.
In this work, we extend the portable Heterogeneous Active Messages (HAM) based HAM-Offload framework with support for the NEC SX-Aurora TSUBASA. Therefore, we design, implement, and evaluate two messaging protocols aimed at minimising offloading cost. This sheds some light on how to achieve fast communication between host CPU and the Vector Engines of the NEC SX-Aurora TSUBASA. Compared with VEO, the DMA-based protocol reduces offloading overhead by a factor of 13×. The resulting framework enables users to write portable offload applications with low overhead, that do neither require a language extension like OpenMP, nor a special language like OpenCL. Existing HAM-Offload applications are now ready to run on the NEC SX-Aurora TSUBASA.
We present HAM (Heterogeneous Active Messages), a C++-only active messaging solution for heterogeneous distributed systems.Combined with a communication protocol, HAM can be used as a generic Remote Procedure Call (RPC) mechanism. It has been used in HAM-Offload to implement a low-overhead offloading framework for inter- and intra-node offloading between different architectures including accelerators like the Intel Xeon Phi x100 series and the NEC SX-Aurora TSUBASA Vector Engine.
HAM uses template meta-programming to implicitly generate active message types and their corresponding handler functions. Heterogeneity is enabled by providing an efficient address translation mechanism between the individual handler code addresses of processes running different binaries on different architectures, as well a hooks to inject serialisation and deserialisation code on a per-type basis. Implementing such a solution in modern C++ sheds some light on the shortcomings and grey areas of the C++ standard when it comes to distributed and heterogeneous environments.
Computing the Hierarchical Equations of Motion (HEOM) is by itself a challenging problem, and so is writing portable production code that runs efficiently on a variety of architectures while scaling from PCs to supercomputers. We combined both challenges to push the boundaries of simulating quantum systems, and to evaluate and improve methodologies for scientific software engineering.
Our contributions are threefold: We present the first distributed memory implementation of the HEOM method (DM-HEOM), we describe an interdisciplinary development workflow, and we provide guidelines and experiences for designing distributed, performance-portable HPC applications with MPI-3, OpenCL and other state-of-the-art programming models. We evaluated the resulting code on multi- and many-core CPUs as well as GPUs, and demonstrate scalability on a Cray XC40 supercomputer for the PS I molecular light harvesting complex.
Performance Evaluation and Improvements of the PoCL Open-Source OpenCL Implementation on Intel CPUs
(2021)
The Portable Computing Language (PoCL) is a vendor independent open-source OpenCL implementation that aims to support a variety of compute devices in a single platform. Evaluating PoCL versus the Intel OpenCL implementation reveals significant performance drawbacks of PoCL on Intel CPUs – which run 92 % of the TOP500 list. Using a selection of benchmarks, we identify and analyse performance issues in PoCL with a focus on scheduling and vectorisation. We propose a new CPU device-driver based on Intel Threading Building Blocks (TBB), and evaluate LLVM with respect to automatic compiler vectorisation across work-items in PoCL. Using the TBB driver, it is possible to narrow the gap to Intel OpenCL and even outperform it by a factor of up to 1.3× in our proxy application benchmark with a manual vectorisation strategy.
Standard offload programming models for the Xeon Phi, e.g. Intel LEO and OpenMP 4.0, are restricted to a single compute node and hence a limited number of coprocessors. Scaling applications across a Xeon Phi cluster/supercomputer thus requires hybrid programming approaches, usually MPI+X. In this work, we present a framework based on heterogeneous active messages (HAM-Offload) that provides the means to offload work to local and remote (co)processors using a unified offload API. Since HAM-Offload provides similar primitives as current local offload frameworks, existing applications can be easily ported to overcome the single-node limitation while keeping the convenient offload programming model. We demonstrate the effectiveness of the framework by using it to enable a real-world application from the field of molecular dynamics to use multiple local and remote Xeon Phis. The evaluation shows good scaling behavior. Compared with LEO, performance is equal for large offloads and significantly better for small offloads.
Effective vectorization is becoming increasingly important for high performance and energy efficiency on processors with wide SIMD units. Compilers often require programmers to identify opportunities for vectorization, using directives to disprove data dependences. The OpenMP 4.x SIMD directives strive to provide portability. We investigate the ability of current compilers (GNU, Clang, and Intel) to generate SIMD code for microbenchmarks that cover common patterns in scientific codes and for two kernels from the VASP and the MOM5/ERGOM application. We explore coding strategies for improving SIMD performance across different compilers and platforms (Intel® Xeon® processor and Intel® Xeon Phi™ (co)processor). We compare OpenMP* 4.x SIMD vectorization with and without vector data types against SIMD intrinsics and C++ SIMD types. Our experiments show that in many cases portable performance can be achieved. All microbenchmarks are available as open source as a reference for programmers and compiler experts to enhance SIMD code generation.