Root/m1/perf/TODO

Source at commit a6136bd02799bcab8d2e48fd21dc91f0fee0df41 created 12 years 5 months ago.
By Werner Almesberger, m1rc3/norruption/2/dumplock: use correct offset; fixed block number counting
1Done:
2- dynamically allocate scheduler context
3- see if preferring critical path can improve code efficiency (YES !)
4
5Pending:
6- see if dynamically adjusting the critical path leads to further improvements
7- test IF
8- run result comparison against full set of patches
9- check if result comparison actually compares meaningful data
10- compare run time and code size for all patches
11- see what optimization changes (may interfere with profiling)
12- build into Flickernoise (some things may need adapting, e.g., abort())
13- review code, see if things can be simplified
14- see if valgrind can do something useful
15

Archive Download this file

Branches:
master



interactive