DGEMM performance is data-dependent

12/11/2019
by   Tom Cornebize, et al.
0

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.

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 1

page 6

page 7

page 9

page 10

05/02/2017

How does Docker affect energy consumption? Evaluating workloads in and out of Docker containers

Context: Virtual machines provide isolation of services at the cost of h...
05/12/2021

Winograd Algorithm for AdderNet

Adder neural network (AdderNet) is a new kind of deep model that replace...
12/20/2021

Fast and Green Computing with Graphics Processing Units for solving Sparse Linear Systems

In this paper, we aim to introduce a new perspective when comparing high...
01/29/2021

The tensor rank of 5x5 matrices multiplication is bounded by 98 and its border rank by 89

We present a non-commutative algorithm for the product of 3x5 by 5x5 mat...
10/26/2018

Comparing Multilayer Perceptron and Multiple Regression Models for Predicting Energy Use in the Balkans

Global demographic and economic changes have a critical impact on the to...
09/04/2019

Engineering Boolean Matrix Multiplication for Multiple-Accelerator Shared-Memory Architectures

We study the problem of multiplying two bit matrices with entries either...
11/22/2018

Building the Case for Temperature Awareness in Energy Consumption Models: an Application of the Energy-Frequency Convexity Rule

Optimizing computing and communication systems that host energy-critical...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.