MMgc's gross stats reporting should include peak memory usage. Would be handy for memory reduction task or OOM investigation.
Where did this come from? I convinced Gavin this isn't necessary as the value can be parsed from GCStats output. Let's close this unless there's a good argument against doing it that way.
Okay. Can you outline how to get that info?
you write a script to parse the stdout when you run your test with -memstats (avmshell) or with the GCStats mm.cfg flag on and you build a table of the private bytes (or whichever) stats and you get the highest one.
Status: NEW → RESOLVED
Last Resolved: 9 years ago
Resolution: --- → WONTFIX
(In reply to Tommy Reilly from comment #1) > Let's close this unless there's a good argument against doing it that way. [Logging some thoughts to myself after I noticed this ticket during a search.] I don't know if this is a "good" argument, but: I've seen AS code that periodically samples System.totalMemory in order to derive the peak value. That to me indicates a desire to know the peak value from within AS3. Also, if we provided peak memory values from Actionscript, our performance benchmarking runtests.py could gather memory and time data in a single run (or rathera, a single series of iterations). Right now the runtests.py gathers memory info via the -memstats option, which from what I've seen disturbs the timing enough that you need to do separate runs for the two. Better still, one might be even able to see potential correlations between time variations and memory usage variations, which is not possible at all currently.
You need to log in before you can comment on or make changes to this bug.