Parsing & Intermediate Files

Processing of input files occurs in several steps:

1. Input Jinja2 templates are parsed and rendered into input YAMLs using the Jinja2 templating functions.
2. YAMLs are parsed into Python objects to be edited by user scripts. See the example tutorials in the Timeloop+Accelergy Exercises.
3. The Python objects are dumped into an intermediate YAML, which is dropped as an input.yaml file as an input to Accelergy+Timeloop.
4. Accelergy uses the intermediate YAML to generate an Energy Reference Table (ERT) and Area Reference Table (ART), which provide energy and area estimates for each component in the architecture.
5. If Timeloop is used, these reference tables will be used to generate full-system energy and area estimates while running a workload.

accessibility