Dear all,
during the development of the 3GPP IVAS codec a couple of issues and potential improvements were identified in the wmc_tool, for example:
- NaN printed in the complexity printout for a non-instrumented function
- update of memory consumption counters could be called from the function
update_wmops()
- suboptimal code updating the worst-case intra-frame and inter-frame heap memory consumption
- detailed per-frame and per-block memory consumption exported to a
.csv file
- incorrect calculation of the worst-case heap size for memory blocks allocated and de-allocated multiple times in a single frame
Exporting the detailed per-frame and per-block memory consumption to an external file allows for running external statistical analysis tools such as Python pandas.
I will provide an update of the wmc_tool addressing all the above issues.
Dear all,
during the development of the 3GPP IVAS codec a couple of issues and potential improvements were identified in the
wmc_tool, for example:update_wmops().csvfileExporting the detailed per-frame and per-block memory consumption to an external file allows for running external statistical analysis tools such as Python pandas.
I will provide an update of the
wmc_tooladdressing all the above issues.