Dataflow dot product on networks of heterogeneous digit-serial arithmetic units

Abstract : In this paper we deal with a new high precision computation of the dot product. The key idea is to use hundreds of digit-serial arithmetic units that allow a massive digit-level pipelining. Parallel discrete-event simulations performed on a memory-distributed massively parallel computer show that with a limited number of arithmetic units, the computation of dot product when performed using a ``classical'' algorithmic technique (i.e. serial cumulative multiplications) is almost as fast as the case where an ``optimal'' divide-and-conquer algorithmic technique is used. Interconnection networks for both algorithmic techniques are considered.
Document type :
Reports
Complete list of metadatas

Cited literature [32 references]  Display  Hide  Download

https://hal-lara.archives-ouvertes.fr/hal-02102063
Contributor : Colette Orange <>
Submitted on : Wednesday, April 17, 2019 - 9:13:19 AM
Last modification on : Friday, May 3, 2019 - 1:30:41 AM

File

RR1993-10.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02102063, version 1

Collections

Citation

Jean Duprat, Mario Fiallos-Aguilar. Dataflow dot product on networks of heterogeneous digit-serial arithmetic units. [Research Report] LIP RR-1993-10, Laboratoire de l'informatique du parallélisme. 1993, 2+17p. ⟨hal-02102063⟩

Share

Metrics

Record views

4

Files downloads

9