Iegulto sistēmu reāllaika profilēšana ar ARM Serial Wire Viewer
Loading...
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Latvijas Universitāte
Language
lav
Abstract
Bakalaura darba ietvaros tika pētīta Advanced RISC Machine (ARM) arhitektūras iegulto sistēmu čipos sastopamās Serial Wire Viewer (SWV) tehnoloģijas piemērotība reāllaika profilēšanai. Profilēšanas metode tika uzskatīta par reāllaika metodi, ja tā būtiski neietekmē profilējamās programmatūras darbību un ja tās rezultātus ir iespējams atjaunot un attēlot profilēšanas procesā. Tika izstrādāta programmatūra, kas izmanto ARM SWV un veikts šīs programmatūras salīdzinājums ar industrijā biežāk izmantotajām profilēšanas metodēm, lai noskaidrotu, vai tās ir iespējams aizvietot ar SWV risinājumu. Salīdzinājumā tika ņemta vērā gan metožu veiktspēja, gan to izmantošanas izmaksas.
A research was conducted with the purpose of having a deeper look at Serial Wire Viewer (SWV) — a technology found in embedded system chips based on Advanced RISC Machine (ARM) architecture, and its suitability for real-time profiling. A profiling method was assumed to be real-time if it doesn’t affect the run-time characteristics of profiled firmware and if its results can be updated and viewed during the profiling process. For testing purposes custom software was built on top of SWV and is compared to frequently used profiling methods in the industry to find out if it’s possible to replace them with SWV–based software. The comparison took into account both the performance of these methods and the monetary cost of using them.
A research was conducted with the purpose of having a deeper look at Serial Wire Viewer (SWV) — a technology found in embedded system chips based on Advanced RISC Machine (ARM) architecture, and its suitability for real-time profiling. A profiling method was assumed to be real-time if it doesn’t affect the run-time characteristics of profiled firmware and if its results can be updated and viewed during the profiling process. For testing purposes custom software was built on top of SWV and is compared to frequently used profiling methods in the industry to find out if it’s possible to replace them with SWV–based software. The comparison took into account both the performance of these methods and the monetary cost of using them.