This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [Patch] Dryrun framework.
- From: OndÅej BÃlka <neleai at seznam dot cz>
- To: Carlos O'Donell <carlos at redhat dot com>
- Cc: libc-alpha at sourceware dot org
- Date: Wed, 10 Apr 2013 10:06:05 +0200
- Subject: Re: [Patch] Dryrun framework.
- References: <20130409192249 dot GA11662 at domone dot kolej dot mff dot cuni dot cz> <516476BC dot 4030605 at redhat dot com>
On Tue, Apr 09, 2013 at 04:14:52PM -0400, Carlos O'Donell wrote:
> On 04/09/2013 03:22 PM, OndÅej BÃlka wrote:
> > Dryrun is framework to supply real world data for microbenchmarks
> > and analyze workloads.
>
> Have you looked into leveraging systemtap to get this
> done with lower overhead and less intrusiveness than
> dlopen()?
>
> With systemptap we add a static probe point to exp()
> with any data that we need.
>
I will add it ass addition in later commit.
Systemtap is limited in various ways, for example that only root can use
it in guru mode.
I plan integrate this with my profiler where dlopen is only sensible way
to go.
> There is a lot of work that has gone into systemtap to
> make sure that it scales and can handle lots of data.
>
> We already have systemtap probe points for the dynamic loader.
> If I ever get around to reviewing Alex's patches we'll have
> systemtap probe points for malloc.
>
> We still need a `dryrun' framework, but in this case
> the framework would help the user turn on the various
> systemtap probe points, collect data in one place, and
> then provide post processing for the results.
>
> Comments?
>
> Cheers,
> Carlos.
--
firewall needs cooling