public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
* gmon size of counter
@ 2004-06-24  9:06 Diviya JAIN
  2004-06-24 13:30 ` Ian Lance Taylor
  0 siblings, 1 reply; 4+ messages in thread
From: Diviya JAIN @ 2004-06-24  9:06 UTC (permalink / raw)
  To: binutils

Hello,

While using gprof, I came across the fact that gmon.h specifies a short counter (kcount) in the gmonparam structure. But we require very large profiling counts, and hence short counters are insufficient for us. 

Is a new version of gmon.h is available which allows for larger counter value, or if you can suggest other possible alternatives that have been implemented till now.

Thank you,
diviya

***************************************************************************
Diviya Jain				Tel(O): 91-120-2515262 Extn-6514
Advanced System Technology Group	TINA: 099-6514
ST Microelectronics, Noida

Modesty is the citadel of beauty - Demades
***************************************************************************


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: gmon size of counter
  2004-06-24  9:06 gmon size of counter Diviya JAIN
@ 2004-06-24 13:30 ` Ian Lance Taylor
  2004-06-24 19:26   ` Andrew Cagney
  0 siblings, 1 reply; 4+ messages in thread
From: Ian Lance Taylor @ 2004-06-24 13:30 UTC (permalink / raw)
  To: Diviya JAIN; +Cc: binutils

Diviya JAIN <diviya.jain@st.com> writes:

> While using gprof, I came across the fact that gmon.h specifies a
> short counter (kcount) in the gmonparam structure. But we require
> very large profiling counts, and hence short counters are
> insufficient for us.

> Is a new version of gmon.h is available which allows for larger
> counter value, or if you can suggest other possible alternatives
> that have been implemented till now.

The structure needs to be kept in synch between the generating
function, which is in glibc, and the reading function, which is in
gprof.

If you change both simultaneously, you can do whatever you like.

Ian

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: gmon size of counter
  2004-06-24 13:30 ` Ian Lance Taylor
@ 2004-06-24 19:26   ` Andrew Cagney
  2004-06-24 19:38     ` Ian Lance Taylor
  0 siblings, 1 reply; 4+ messages in thread
From: Andrew Cagney @ 2004-06-24 19:26 UTC (permalink / raw)
  To: Ian Lance Taylor; +Cc: Diviya JAIN, binutils

> Diviya JAIN <diviya.jain@st.com> writes:
> 
> 
>>> While using gprof, I came across the fact that gmon.h specifies a
>>> short counter (kcount) in the gmonparam structure. But we require
>>> very large profiling counts, and hence short counters are
>>> insufficient for us.
> 
> 
>>> Is a new version of gmon.h is available which allows for larger
>>> counter value, or if you can suggest other possible alternatives
>>> that have been implemented till now.
> 
> 
> The structure needs to be kept in synch between the generating
> function, which is in glibc, and the reading function, which is in
> gprof.
> 
> If you change both simultaneously, you can do whatever you like.

Is there a way to check for counter-size consistency between gprof and 
glibc?

Andrew


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: gmon size of counter
  2004-06-24 19:26   ` Andrew Cagney
@ 2004-06-24 19:38     ` Ian Lance Taylor
  0 siblings, 0 replies; 4+ messages in thread
From: Ian Lance Taylor @ 2004-06-24 19:38 UTC (permalink / raw)
  To: Andrew Cagney; +Cc: Diviya JAIN, binutils

Andrew Cagney <cagney@gnu.org> writes:

> > Diviya JAIN <diviya.jain@st.com> writes:
> >
> >>> While using gprof, I came across the fact that gmon.h specifies a
> >>> short counter (kcount) in the gmonparam structure. But we require
> >>> very large profiling counts, and hence short counters are
> >>> insufficient for us.
> >
> >>> Is a new version of gmon.h is available which allows for larger
> >>> counter value, or if you can suggest other possible alternatives
> >>> that have been implemented till now.
> > The structure needs to be kept in synch between the generating
> > function, which is in glibc, and the reading function, which is in
> > gprof.
> > If you change both simultaneously, you can do whatever you like.
> 
> Is there a way to check for counter-size consistency between gprof and
> glibc?

There is a magic number.  I don't think there is anything which
explicitly states the size of the histogram entries.

(They are 16 bits in all existing implementations.)

Ian

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2004-06-24 19:38 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-06-24  9:06 gmon size of counter Diviya JAIN
2004-06-24 13:30 ` Ian Lance Taylor
2004-06-24 19:26   ` Andrew Cagney
2004-06-24 19:38     ` Ian Lance Taylor

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).