Re: [tree-ssa mudflap] slow startup

["Frank Ch. Eigler" <fche at redhat dot com>]

Hi -

On Sun, Jun 29, 2003 at 03:41:31PM +1000, Eyal Lebedinsky wrote:
> [...]
> Following up on my own post, looking at a trace of my simple
> program with the old and then the new compiler, I see that there
> is a significant increase of 'string literal's registered. [...]

This is probably more fallout from the removal of the TREE_MUDFLAPPED
(a hijacked TREE_BOUNDED) tree node flag.  I can't think of a reason
that mudflap instrumentation-generated string literals (and there
are lots) should be included in the libmudflap database.  I suspect
those strings are for some reason not being flagged.

As for tree geometry, you're right that the balancing algorithm should
be smarter when constructing of the tree without liveness
(frequently-accessed-status) statistics.  Improvements are welcome :-)
but once the above problem is dealt with, it may not be crucial.

As for multithreading (reported last week), I came across a big fat
problem in libmudflap thread support which I've been working on.
(The "reentrant" indication must be per-thread instead of global,
among other problems.)


- FChE

Attachment: pgp00000.pgp
Description: PGP signature