Abstract
This paper focuses on challenging applications that can be expressed as an iterative pipeline of multiple 3d stencil stages and explores their optimization space on GPUs. For this study, we selected a representative example from the field of digital signal processing, the Anisotropic Nonlinear Diffusion algorithm. An open issue to these applications is to determine the optimal fission/fusion level of the involved stages and whether that combination benefits from data tiling. This implies exploring a large space of all the possible fission/fusion combinations with and without tiling, thus making the process non-trivial. This study provides insights to reduce the optimization tuning space and programming effort of iterative multiple 3d stencils. Our results demonstrate that all combinations that fuse the bottleneck stencil with high halos update cost (> 25 % , this percentage can be measured or estimated experimentally for each single stencil) and high registers and shared memory accesses must not be considered in the exploration process. The optimal fission/fusion combination is up to 1.65× faster than the case in which we fully decompose our stencil without tiling and 5.3× faster with respect to the fully fused version on the NVIDIA GPUs.
Original language | English |
---|---|
Pages (from-to) | 1580-1608 |
Number of pages | 29 |
Journal | Journal of Supercomputing |
Volume | 74 |
Issue number | 4 |
DOIs | |
Publication status | Published - 1 Apr 2018 |
Fingerprint
Keywords
- 3d images
- 3d stencils
- Anisotropic Nonlinear Diffusion
- Fission
- Fusion
- GPUs
- Tiling
Cite this
}
A tuning approach for iterative multiple 3d stencil pipeline on GPUs : Anisotropic Nonlinear Diffusion algorithm as case study. / Tabik, S.; Peemen, M.; Romero, L. F.
In: Journal of Supercomputing, Vol. 74, No. 4, 01.04.2018, p. 1580-1608.Research output: Contribution to journal › Article › Academic › peer-review
TY - JOUR
T1 - A tuning approach for iterative multiple 3d stencil pipeline on GPUs
T2 - Anisotropic Nonlinear Diffusion algorithm as case study
AU - Tabik, S.
AU - Peemen, M.
AU - Romero, L. F.
PY - 2018/4/1
Y1 - 2018/4/1
N2 - This paper focuses on challenging applications that can be expressed as an iterative pipeline of multiple 3d stencil stages and explores their optimization space on GPUs. For this study, we selected a representative example from the field of digital signal processing, the Anisotropic Nonlinear Diffusion algorithm. An open issue to these applications is to determine the optimal fission/fusion level of the involved stages and whether that combination benefits from data tiling. This implies exploring a large space of all the possible fission/fusion combinations with and without tiling, thus making the process non-trivial. This study provides insights to reduce the optimization tuning space and programming effort of iterative multiple 3d stencils. Our results demonstrate that all combinations that fuse the bottleneck stencil with high halos update cost (> 25 % , this percentage can be measured or estimated experimentally for each single stencil) and high registers and shared memory accesses must not be considered in the exploration process. The optimal fission/fusion combination is up to 1.65× faster than the case in which we fully decompose our stencil without tiling and 5.3× faster with respect to the fully fused version on the NVIDIA GPUs.
AB - This paper focuses on challenging applications that can be expressed as an iterative pipeline of multiple 3d stencil stages and explores their optimization space on GPUs. For this study, we selected a representative example from the field of digital signal processing, the Anisotropic Nonlinear Diffusion algorithm. An open issue to these applications is to determine the optimal fission/fusion level of the involved stages and whether that combination benefits from data tiling. This implies exploring a large space of all the possible fission/fusion combinations with and without tiling, thus making the process non-trivial. This study provides insights to reduce the optimization tuning space and programming effort of iterative multiple 3d stencils. Our results demonstrate that all combinations that fuse the bottleneck stencil with high halos update cost (> 25 % , this percentage can be measured or estimated experimentally for each single stencil) and high registers and shared memory accesses must not be considered in the exploration process. The optimal fission/fusion combination is up to 1.65× faster than the case in which we fully decompose our stencil without tiling and 5.3× faster with respect to the fully fused version on the NVIDIA GPUs.
KW - 3d images
KW - 3d stencils
KW - Anisotropic Nonlinear Diffusion
KW - Fission
KW - Fusion
KW - GPUs
KW - Tiling
UR - http://www.scopus.com/inward/record.url?scp=85033433676&partnerID=8YFLogxK
U2 - 10.1007/s11227-017-2184-6
DO - 10.1007/s11227-017-2184-6
M3 - Article
AN - SCOPUS:85033433676
VL - 74
SP - 1580
EP - 1608
JO - Journal of Supercomputing
JF - Journal of Supercomputing
SN - 0920-8542
IS - 4
ER -