Oak Ridge National Laboratory

By L. Love, Ellen Mell. 2002

— The Rodinia Benchmark Suite in SYCL was first published as a technical report at Argonne National Laboratory in June 2020. While porting more programs to SYCL, the Rodinia benchmark suites were merged into the heterogenous computing benchmarks (HeCBench). Since then, the source codes and scripts of the benchmark suite have been updated to fix compile-and run-time issues that users encountered. With the recent SYCL compilers from Codeplay and Intel, this report presents the experimental results of evaluating these benchmarks on the Intel CPU and GPU devices.

Subfields: ML Systems & Hardware

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In ML Systems & Hardware

1. Sebastian Kriticos
2. Dzhamilya Nigmatulina
3. J. Henderson

Citations in the corpus

• ZeRO-Infinity: Breaking the GPU Memory Wall for Extreme Scale Deep learning (2021, methodology): “For example, the Summit Supercomputer launched in 2018 [45] connects NVIDIA V100 GPUs with the CPU memory at 40GB/s per GPU.”
• Measuring urban economic density (2018, methodology): “Section B.1 and its sub-parts describe the LandScan (2012) data and how we use it to construct urbanized area boundaries, as well as why we chose certain density cut-off criteria for the construction of the boundaries.”