You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Regression tests today launch Timeloop using subprocess and extract results by parsing XML stats dumped by Timeloop using Boost. Porting regression tests to use TimeloopFE interfaces will clean up the interface and remove the Boost dependency, allowing us to remove Boost from the build process completely. @tanner-andrulis
The text was updated successfully, but these errors were encountered:
Simplifying the interface and dropping boost would be great. A couple questions:
What do these regression tests compare? The TimeloopFE parser only picks up high-level stats like per-component energy, per-component area, overall area/energy/throughput/latency, and mapping. It doesn't grab the exact access counts for every component however.
How would you feel about expanding them to use the accelerators in https://github.com/Accelergy-Project/timeloop-accelergy-exercises? There are already several in there (including Simba). I think a good test would be to run all the example designs and check for differences in the output.
Regression tests today launch Timeloop using subprocess and extract results by parsing XML stats dumped by Timeloop using Boost. Porting regression tests to use TimeloopFE interfaces will clean up the interface and remove the Boost dependency, allowing us to remove Boost from the build process completely. @tanner-andrulis
The text was updated successfully, but these errors were encountered: