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
For embench-iot 2.0 we discussed that we would like to modernize/update the python build scripts. @wallento wanted to take a look at it if I'm correct.
Some things that we could address:
change the --target-module speed benchmarking architecture.
It might be better to use callbacks/a more flexible approach that doesn't rely on a user-constructed shell command.
use a more declarative framework such as doit
reduce usage of global data/variables
update the benchmark size computation
right now, the size is computed via hardcoded section names, however, some sections aren't accounted for: e.g. .eh_frame, .got, .init, .plt, maybe we should instead sum up all MemSiz fields of all segments that get loaded?
For embench-iot 2.0 we discussed that we would like to modernize/update the python build scripts.
@wallento wanted to take a look at it if I'm correct.
Some things that we could address:
--target-modulespeed benchmarking architecture..eh_frame, .got, .init, .plt, maybe we should instead sum up all MemSiz fields of all segments that get loaded?