OmpSs check benchmarks

This is a collection of benchmarks using the OmpSs programming model that include a verification of their results, either algorithmical or comparing with a reference output. These codes are provided without guarantee of any kind.

Code organisation

Each directory provides build targets perf, instr, debug, seq, install, uninstall. The first three are parallel builds, respectively built with and linked against the performance, instrumentation, and debug libraries of the Nanox++ runtime. seq builds a sequential binary. All builds require a built libcatchroi at $CATCHROI_HOME (defaults to <repo_root>/libcatchroi). install and uninstall targets respect the $DESTDIR variable.

This repository contains two libraries:

• libcatchroi, that provides timing information and some error injection capabilities. The library format is useful to interpose the Region Of Interest (ROI) start and end calls, as well as any desired memory allocation calls.
• nx_catch_tdg, which is a nanox instrumentation plugin to report scheduling information.

The rest of the subdirectories are benchmarks.

Benchmarks used for evaluation

The table below lists all the benchmarks in this repository. The "Verif." column indicates how the benchmark's output verification is performed.

Name	Benchmark description	Category	Verif	Origin
Blackscholes	Option pricing	Partial Differential Equation	built-in	PARSEC benchmarks4,5
Cholesky	Cholesky factorization	Dense linear algebra	built-in	BSC Application Repository
CG	Conjugate Gradient	Sparse linear algebra	built-in	matrices from SuiteSparse2
DGEMM	Matrix multiplication	Dense linear algebra	built-in	BSC Application Repository
FFT	Fast Fourier Transform	Spectral method	ref. run	Wang Jian-Sheng
Gauss-Seidel	Heat diffusion, Gauss-Seidel solver	Structured grid	ref. run	BSC Application Repository
Jacobi	Heat diffusion, Jacobi solver	Structured grid	ref. run	BSC Application Repository
KNN	K-nearest neighbours	Machine learning	ref. run	Heterogeneous Computer Architecture (HCA) group at BSC
K Means	K-means clustering	Machine learning	ref. run	Heterogeneous Computer Architecture (HCA) group at BSC
N-body	Astrophysical simulation	N-body method	built-in	BSC Application Repository
PRK2 stencil	Parallel Research Kernels stencil	Stencil operation	built-in	Parallel Research Kernels3
Red-black	Heat diffusion, red-black solver	Structured grid	ref. run	BSC Application Repository
SMI	Symmetric matrix inverse	Dense linear algebra	built-in	Guillermo Miranda
Stream	Stream Triad	Memory bandwidth benchmark	built-in	John D. McCalpin

The BSC benchmarks repository was previously hosted at https://pm.bsc.es/projects/bar. SuiteSparse was previously the UF Sparse Matrix Collection hosted at https://www.cise.ufl.edu/research/sparse/matrices/list_by_id.html.

1. J. D. McCalpin, “Memory Bandwidth and Machine Balance in Current High Performance Computers,” IEEE Computer Society Technical Committee on Computer Architecture (TCCA) Newsletter, pp. 19–25, 1995.
2. T. A. Davis and Y. Hu, “The University of Florida Sparse Matrix Collection,” ACM Transactions on Mathematical Software, vol. 38, no. 1, pp. 1:1–1:25, Dec. 2011.
3. R. F. V. der Wijngaart and T. G. Mattson, “The Parallel Research Kernels,” in IEEE High Performance Extreme Computing Conference, 2014, pp. 1–6.
4. C. Bienia, “Benchmarking Modern Multiprocessors,” PhD thesis, Princeton University, 2011.
5. D. Chasapis et al., “PARSECSs: Evaluating the Impact of Task Parallelism in the PARSEC Benchmark Suite,” ACM Transactions on Architecture and Code Optimization, vol. 12, no. 4, pp. 41:1–41:22, Dec. 2015.

Licensing

All codes are under their original license. This repository does not own any of the code and does not provide any warranties. Code written that is not part of any benchmark (under nx_catch_tdg/, libcatchroi/) is under LGPL v3.