[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Botan-devel] Re: [Monotone-devel] oprofile data for mtn 0.37.
From: |
Jack Lloyd |
Subject: |
Re: [Botan-devel] Re: [Monotone-devel] oprofile data for mtn 0.37. |
Date: |
Mon, 7 Apr 2008 10:20:28 -0400 |
User-agent: |
Mutt/1.5.11 |
On Wed, Apr 02, 2008 at 03:14:58PM -0400, Jack Lloyd wrote:
> On Wed, Mar 12, 2008 at 05:19:50PM -0400, Zack Weinberg wrote:
> > > > Top of profile is
> > > >
> > > > 9678874 11.1247 std::_Rb_tree<std::string, std::pair<std::string
> > > > const, Botan::Mutex*> ...>
> >
> [...]
> > > 3) Drop global shared state wholesale, and require applications to
> > > maintain any per-thread or per-process state (the RNG, allocators,
> > > etc). [I've been leaning towards this of late for various other
> > > reasons anyway]
> >
> > This would be my choice, being the most efficient option for a
> > single-threaded program like Monotone.
> >
> > zw
>
> As a start in this direction (removing the global state will be a
> fairly long/incremental process), and to address the immediate
> performance concern, I've checked in a revision to n.r.b
> (902a87e6d79771db7a64cc115b279febb4c936ee) that removes the
> string->mutex mapping code and replaces it with plain objects.
I'm checking in changes for incremental removal of code from the
global state object into net.randombit.botan.remove-libstate
-Jack