This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFA/commit/dwarf] Create partial symbols for nested subprograms


> Joel> It looks like roughly a 4% increase in startup time. Not sure whether
> Joel> that's considered a large increase or not - I just think that it's not
> Joel> noticeable. None of our users have reported issues with startup time.
> 
> How much memory do you have on your machine?  Are you using a 32- or
> 64-bit machine?  And how many objfiles are made with this test
> program?

Phew, all these technical questions :). I'm using a 32bit machine
(Intel(R) Core(TM)2 CPU T7400 @ 2.16GHz) with 2GB of memory.
I'm running a 2.6.19.3 Linux kernel.

Because I cannot run the test program (all these large programs that
we got from our customers require a very precise environement to start),
I can only time "gdb EXE". As a result, I think we only end up creating
one object file for that executable.

> I am just curious to know what differs between my tests and yours.  In
> my case, I start the system OpenOffice writer and then attach to it.

Yes, that's a little different, because the shared libraries have been
mapped, and so debugging info for these are loaded in as well.

> FWIW I don't think Joel's patch will negatively affect my current
> approach.

Just curious, I think I either forgot a previous discussion or missed
it entirely. What approach are you going to take?

-- 
Joel


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]