polybench-tiramisu

January 29, 2021 · View on GitHub

PolyBench-Tiramisu is the reimplementation of PolyBench 4.2 in the Tiramisu programming language. It is a benchmark suite of 30 numerical computations with static control flow, extracted from operations in various application domains (linear algebra computations, image processing, physics simulation, dynamic programming, statistics, etc.).

Usage

Set the environment variable TIRAMISU_ROOT to the root directory of your Tiramisu installation.

export TIRAMISU_ROOT=<path/to/Tiramisu>

Run compile_and_run.sh specifying the benchmark name and the problem size.

./compile_and_run.sh <benchmark name> <problem size>

Example:

./compile_and_run.sh cholesky LARGE

Available Benchmarks

BenchmarkDescription
2mm2 Matrix Multiplications (E=A.B; F=E.C; G=F+D)
3mm3 Matrix Multiplications (E=A.B; F=C.D; G=E.F)
adiAlternating Direction Implicit solver
ataxMatrix Transpose and Vector Multiplication
bicgBiCG Sub Kernel of BiCGStab Linear Solver
choleskyCholesky Decomposition
correlationCorrelation Computation
covarianceCovariance Computation
dericheEdge detection filter
doitgenMulti-resolution analysis kernel (MADNESS)
durbinToeplitz system solver
fdtd_2d2-D Finite Different Time Domain Kernel
gemmMatrix-multiply C=alpha.A.B+beta.C
gemverVector Multiplication and Matrix Addition
gesummvScalar, Vector and Matrix Multiplication
gramschmidtGram-Schmidt decomposition
head-3dHeat equation over 3D data domain
jacobi_1D1-D Jacobi stencil computation
jacobi_2D2-D Jacobi stencil computation
luLU decomposition
ludcmpLU decomposition followed by FS
mvtMatrix Vector Product and Transpose
nussinovDynamic programming algorithm for sequence alignment
seidel_2d2-D Seidel stencil computation
symmSymmetric matrix-multiply
syr2kSymmetric rank-2k operations
syrkSymmetric rank-k operations
trisolvTriangular solver
trmmTriangular matrix-multiply

Available Sizes

Problem sizeMemory usage
MINI< 16KB of memory.
SMALL≈ 128KB of memory.
MEDIUM≈ 1MB of memory.
LARGE≈ 25MB of memory.
EXTRALARGE≈ 120MB of memory.

References