https://bugs.kde.org/show_bug.cgi?id=402369
Philippe Waroquiers <philippe.waroqui...@skynet.be> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |philippe.waroquiers@skynet.
| |be
--- Comment #3 from Philippe Waroquiers <philippe.waroqui...@skynet.be> ---
No comment about the tool itself, just a suggestion for the implementation.
As I understand, dhat has 2 data structures:
* an interval tree, to find for an address a what block is covering it
* a wordFM of an Execontext, together with some data associated with
this execontext.
It looks to me that the second data structure could be done by using
pub_tool_xtree.h, which is a generalised way to associate data to an
execontext, and print it in (currently) 2 output formats : massif format,
kcachegrind format.
It might be interesting to replace the wordFM by an xtree,
and add the json output format to pub_tool_xtree.h
(I suspect that an xtree might be somewhat faster than an wordFM).
Massif speed was improved a lot when switching it to an xtree.
--
You are receiving this mail because:
You are watching all bug changes.
