Vectorised Computation of Diverging Ensembles. In: ICPP 2018 Proceedings of the 47th International Conference on Parallel Processing, Article No. 51

Conference · Mon Jan 01 00:00:00 EST 2018

DOI:https://doi.org/10.1145/3225058.3225138· OSTI ID:1567654

Hückelheim, Jan ^[1]; Hovland, Paul ^[2]; Narayanan, Sri Hari Krishna ^[2]; Velesko, Paulius ^[3]

---

1. Imperial College London, London
2. Argonne National Laboratory, Argonne, Illinois
3. Intel Corporation, Argonne, Illinois

Ensemble computations are used to evaluate a function for multiple inputs, for example in uncertainty quantification. Embedded ensemble computations perform several evaluations within the same program, often enabling a reduced overall runtime by exploiting vectorisation and parallelisation opportunities that are not present in individual ensemble members. This is challenging if members take different control flow paths. We present a source-to-source transformation that turns a given C program into an embedded ensemble program that computes members in a single-instruction-multiple-data fashion using OpenMP SIMD pragmas. We use techniques from whole-function vectorisation, achieving effective vectorisation for moderate amounts of branch divergence, particularly on processors with masked instructions such as recent Xeon Phi or Skylake processors with AVX-512.

• Research Organization: Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States). Oak Ridge Leadership Computing Facility (OLCF)
• Sponsoring Organization: USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
• DOE Contract Number: AC02-06CH11357
• OSTI ID: 1567654
• Resource Relation: Conference: 47th International Conference on Parallel Processing, Eugene, OR, USA, August 13 - 16, 2018
• Country of Publication: United States
• Language: English