stats: move code that loops over all stats into python
Review Request #690 - Created May 9, 2011 and submitted
| Information | |
|---|---|
| Nathan Binkert | |
| gem5 | |
| Reviewers | |
| Default | |
| ali, gblack, nate, stever | |
stats: move code that loops over all stats into python
quick regressions pass (though most recently run with review 689 and 691)
Posted (May 10, 2011, 8:24 a.m.)
-
src/base/stats/text.hh (Diff revision 1) -
What do begin and end() do?
-
src/base/stats/text.cc (Diff revision 1) -
Should this still be printed for here? Or rather, what is the point of printing some text here and all the stats from python?
-
src/base/stats/text.cc (Diff revision 1) -
Same question here? Why not just create this in python?
-
src/python/m5/stats/__init__.py (Diff revision 1) -
In what cases does dump() get called twice?
Posted (May 11, 2011, 1:24 p.m.)
I guess I should have said something, but there are several more diffs to come down the road. I just wanted to get some out of the way now. By the end, there will be no output code in C++ at all.
-
src/python/m5/stats/__init__.py (Diff revision 1) -
If you have a periodic dumping of stats because of an event and you get a dumpstats from an m5op or from kill -USR1, you don't want to dump twice in the same cycle because stuff just doesn't work well.