|
|
|
Re: Finding top n largest strings [message #3937 is a reply to message #3913] |
Mon, 01 September 2008 09:26 |
Andreas Buchen Messages: 123 Registered: July 2009 |
Senior Member |
|
|
Hi tt :-)
> I seems like I have a heap dump (with indexes) for which the histogram
> calculation is extremely slow (still running after 15 minutes).
That is strange. The default histogram is not really calculated, but
should always be available right away. The default histogram is the one
that is opened from the toolbar button.
What could be slow, though, is the calculation of the retained sizes for
each line of the histogram. Use the approximate times instead. It has an
error of about 2 to 4 percent, but it is ways faster. Any you'll find the
big chunks either way. But this functionality has to be triggered from the
toolbar separately.
Also, the histogram is used in other contexts too (e.g. displaying subsets
like all objects kept alive by all hash maps etc.). Depending on the
context, the calculation could take some time. But not for the default
histogram.
> Out of curiosity, are the index files used in a read-only manner?
Yes, the index files are created during the first parsing. After that,
they are used read-only. You could transfer them to another machine along
the heap dump and reopen the heap dump there.
Andreas.
|
|
|
Powered by
FUDForum. Page generated in 0.02464 seconds