mi1 vs mi3 usage

[lauruhl]

i did a test, in order to understand the difference between those 2 options. i reran the same test, changing in the GUI from mi1 to mi3 and the funny thing is that when using mi3 , i was able to detect much more memoery leaks ... i wonder why . i would hve thought that mi1 would be able to detect all memory leaks that mi3 detected ....
does anybody has a clue ?
thx ,
Laurent

[Peter_W_Intel]

In general speaking, "mi3" option does more work than "mi1" - for example, "mi1" detects memory leas for "Stack frame depth" as 8, but "mi3" does it for "Stack frame depth" as 16. It means that "mi3" will do more work incallers (functions), mostlymemory leaks are detectedin system libraries.

You might use Project Properties -> Target -> Advance -> Modules: Include only the following module(s), then click "Modify..." button to input your application. Thus, other modules will not be checked, "mi1" and "mi3" options will has similar report (note that "mi3" still does "dangling pointer check", and other invalid memory access but "mi1" does NOT).

Regards, Peter