SYCL-Bench: A Versatile Cross-Platform Benchmark Suite for Heterogeneous Computing

Abstract

The SYCL standard promises to enable high productivity in heterogeneous programming of a broad range of parallel devices, including multicore CPUs, GPUs, and FPGAs. Its modern and expressive C++ API design, as well as flexible task graph execution model give rise to ample optimization opportunities at run-time, such as the overlapping of data transfers and kernel execution. However, it is not clear which of the existing SYCL implementations perform such scheduling optimizations, and to what extent. Furthermore, SYCL's high level of abstraction may raise concerns about sacrificing performance for ease of use. Benchmarks are required to accurately assess the performance behavior of high-level programming models such as SYCL. To this end, we present SYCL-Bench, a versatile benchmark suite for device characterization and runtime benchmarking, written in SYCL. We experimentally demonstrate the effectiveness of SYCL-Bench by performing device characterization of the NVIDIA TITAN X GPU, and by evaluating the efficiency of the hipSYCL and ComputeCpp SYCL implementations.

Benchmarking and Extending SYCL Hierarchical Parallelism

“…On CPU architectures, we find the challenges of NDRange parallelism in library-only SYCL implementations identified by Lal et al [4], as seen in Figures 2a and 2b. With compiler support, these problems are alleviated as shown in Figure 2c.…”

“…This can add considerable overhead and limit vectorization across work-items in library-only SYCL implementations if explicit barriers or collective group algorithms are used (see e.g. [4] for details). However, mapping nd_range parallelism to e.g.…”

“…The SYCL Benchmark suite (SYCL-Bench) contains a collect of idiomatic kernels implemented in SYCL for microbenchmarking, testing SYCL runtime overheads and proxies of applications [4], [8]. We select three kernels for our study: nbody, reduction and segmented reduction.…”

SYCL in the Edge: Performance Evaluation for Heterogeneous Acceleration

“…SYCL-Bench also integrates fifteen kernels/applications from Polybench. This suite also has the possibility to execute on different SYCL implementations, such as DPC++, ComputeCpp, triSYCL, and AdaptiveCpp [31].…”

“…The Atom CPU with the AdaptiveCpp compiler obtains the worst performance because SYCL code is translated to OpenMP, while DPC++ is conducted by the OpenCL backend. This point makes the difference between both implementations [31,50]. When comparing DPC++ performance on both CPU and GPU (UHD Graphics), it's noteworthy that the Atom processor even outperforms the GPU.…”