In recent years, Intel introduced oneAPI as a unified and cross-architecture programming model based on the Data Parallel C++ (DPC++) language, which in turn, is based on the C++ and SYCL standard languages. In order to facilitate the migration of legacy CUDA code originally written for NVIDIA GPUs, developers can employ the Intel DPC++ Compatibility Tool, which aims to automatically migrate code from CUDA to SYCL. While this tool-assisted code migration is a good starting point for leveraging the Intel oneAPI ecosystem, manual steps for code completion and tuning are still required. In this paper, we present our experiences migrating AutoDock-GPU, a widely-used molecular docking application, from CUDA to SYCL. Our discussion focuses on: (1) the use of this automated source-code migration tool, (2) the required manual code refinement for functionality and optimization, and (3) the comparison of the performance achieved in this manner on multi-core CPUs as well as on high-end GPUs, such as NVIDIA A100 and the recently-launched Intel Data Center Max 1550 device.
In this work, we introduce Altis-SYCL, a benchmark suite based on SYCL for GPUs and FPGAs. For developing Altis-SYCL, we leverage the oneAPI heterogeneous programming framework in two consecutive steps: 1) by using the modern Altis GPGPU benchmark suite as baseline and migrating it from CUDA to SYCL, and 2) by exploring several techniques to optimize the performance of the resulting SYCL code. Our migration-and-optimization methodology starts targeting GPUs and progressively moves towards FPGAs. In this process, we discuss the differences between device-specific strategies as well as detailing the required code refactoring and optimization efforts. The performance of Altis-SYCL was evaluated on Stratix 10 and Agilex FPGAs, and for some applications, their execution runtimes were competitive with those achieved on latest high-end GPUs. The corresponding code is released as open source under: https://github.com/esa-tu-darmstadt/altis_sycl.
16 March 2021 in Distributed and Parallel Databases
Massive data transfers in modern data-intensive systems resulting from low data-locality and data-to-code system design hurt their performance and scalability. Near-Data processing (NDP) and a shift to code-to-data designs may represent a viable solution as packaging combinations of storage and compute elements on the same device has become feasible. The shift towards NDP system architectures calls for revision of established principles. Abstractions such as data formats and layouts typically spread multiple layers in traditional DBMS, the way they are processed is encapsulated within these layers of abstraction. The NDP-style processing requires an explicit definition of cross-layer data formats and accessors to ensure in-situ executions optimally utilizing the properties of the underlying NDP storage and compute elements. In this paper, we make the case for such data format definitions and investigate the performance benefits under RocksDB and the COSMOS hardware platform.