Abstract
This work presents an early stage statement-level metric for energy characterization of embedded processors. Definition and the framework for metric evaluation are provided. In particular, such a metric is based on an existing assembly-level analysis and some profiling activities performed on a given C benchmark, and it is related to the average energy consumption of a generic C statement, for a given target processor. Its evaluation is performed with a one-time effort and, once available, it can be used to rapidly estimate the energy consumption of a given C function for all the considered processors. Two reference embedded processors are then considered in order to show an example of usage of the proposed metric and framework.
| Original language | English |
|---|---|
| Article number | 103200 |
| Pages (from-to) | 1-18 |
| Number of pages | 18 |
| Journal | Microprocessors and Microsystems |
| Volume | 77 |
| Early online date | 8 Jul 2020 |
| DOIs | |
| Publication status | Published - Sept 2020 |
Funding
This work has been partially supported by the ECSEL RIA 2017-783162 FitOptiVis and ECSEL RIA 2018-826610 COMP4DRONES European projects.
| Funders | Funder number |
|---|---|
| Horizon 2020 Framework Programme | 826610, 783162 |
| Electronic Components and Systems for European Leadership | RIA 2018-826610 |
Fingerprint
Dive into the research topics of 'An early-stage statement-level metric for energy characterization of embedded processors'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver