We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.DC

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Distributed, Parallel, and Cluster Computing

Title: A Technique for Finding Optimal Program Launch Parameters Targeting Manycore Accelerators

Abstract: In this paper, we present a new technique to dynamically determine the values of program parameters in order to optimize the performance of a multithreaded program P. To be precise, we describe a novel technique to statically build another program, say, R, that can dynamically determine the optimal values of program parameters to yield the best program performance for P given values for its data and hardware parameters. While this technique can be applied to parallel programs in general, we are particularly interested in programs targeting manycore accelerators. Our technique has successfully been employed for GPU kernels using the MWP-CWP performance model for CUDA.
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF)
Cite as: arXiv:1906.00142 [cs.DC]
  (or arXiv:1906.00142v1 [cs.DC] for this version)

Submission history

From: Davood Mohajerani [view email]
[v1] Sat, 1 Jun 2019 03:32:57 GMT (293kb)

Link back to: arXiv, form interface, contact.