DGEMM performance is data-dependent

Abstract : The DGEMM function is a widely used implementation of the matrix product. While the asymptotic complexity of the algorithm only depends on the sizes of the matrices, we show that the performance is significantly impacted by the matrices content. Our experiments show that this may be due to bit flips in the CPU causing an energy consumption overhead.
Keywords : Performance BLAS DVFS DGEMM
Complete list of metadatas

https://hal.inria.fr/hal-02401760
Contributor : Tom Cornebize <>
Submitted on : Wednesday, December 11, 2019 - 4:08:15 PM
Last modification on : Friday, December 27, 2019 - 3:20:27 PM

Files

RR-9310.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02401760, version 1
  • ARXIV : 1912.05381

Citation

Tom Cornebize, Arnaud Legrand. DGEMM performance is data-dependent. [Research Report] RR-9310, Université Grenoble Alpes; Inria; CNRS. 2019. ⟨hal-02401760⟩

Share

Metrics

Record views

142

Files downloads

212