[ECOS] Timing with longjmp

[ECOS] Timing with longjmp

[Michele Portolan]

Is there a test in ecos2.0 release to measure the execution time needed by
"longjmp"?
If not, hwo can I do it (apart from counting assembler lines)?

Michele


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] Timing with longjmp

[Andrew Lunn]

On Thu, Aug 28, 2003 at 03:58:47PM +0200, Michele Portolan wrote:
> Is there a test in ecos2.0 release to measure the execution time needed by
> "longjmp"?

Not that i know off, but i've not gone looking.

> If not, how can I do it (apart from counting assembler lines)?

Use the ideas in tm_basic. It has the basic infrastructure to do the
measurements, you just need to write a test inside the framework.

        Andrew 

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] Timing with longjmp

[Gary Thomas]

On Thu, 2003-08-28 at 07:58, Michele Portolan wrote:
> Is there a test in ecos2.0 release to measure the execution time needed by
> "longjmp"?
> If not, hwo can I do it (apart from counting assembler lines)?

There is not currently a test to time this.  If you look at the
code in 'tm_basic.cxx', you can get a pretty good idea how to measure
such [small] quantities.  Basically, you need to capture the current
raw clock time (tm_basic uses 'HAL_CLOCK_READ()' for this).  Grab the
value before doing the longjump and then after you land in the new
location.  Then it's just simple math to figure out how long it took.
Again, tm_basic can show you how to convert the values to real time
(micro-seconds or whatever).

Why are you bothering with such minutia?

-- 
Gary Thomas <gary@mlbassoc.com>
MLB Associates


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

[ECOS] R: [ECOS] Timing with longjmp

[Michele Portolan]

> There is not currently a test to time this.  If you look at the
> code in 'tm_basic.cxx', you can get a pretty good idea how to measure
> such [small] quantities.  Basically, you need to capture the current
> raw clock time (tm_basic uses 'HAL_CLOCK_READ()' for this).  Grab the
> value before doing the longjump and then after you land in the new
> location.  Then it's just simple math to figure out how long it took.
> Again, tm_basic can show you how to convert the values to real time
> (micro-seconds or whatever).
>
> Why are you bothering with such minutia?

I am making some peculiar timing measures on eCos to develop some idea, as I
said on a precedent post. That is the reason why I am so interested in such
little things: they could make the difference.

Michele


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] R: [ECOS] Timing with longjmp

[Jonathan Larmour]

Michele Portolan wrote:
>>There is not currently a test to time this.  If you look at the
>>code in 'tm_basic.cxx', you can get a pretty good idea how to measure
>>such [small] quantities.  Basically, you need to capture the current
>>raw clock time (tm_basic uses 'HAL_CLOCK_READ()' for this).  Grab the
>>value before doing the longjump and then after you land in the new
>>location.  Then it's just simple math to figure out how long it took.
>>Again, tm_basic can show you how to convert the values to real time
>>(micro-seconds or whatever).
>>
>>Why are you bothering with such minutia?
> 
> I am making some peculiar timing measures on eCos to develop some idea, as I
> said on a precedent post. That is the reason why I am so interested in such
> little things: they could make the difference.

Watch out for HAL_CLOCK_READ wrapping BTW.

Jifl
-- 
eCosCentric    http://www.eCosCentric.com/    The eCos and RedBoot experts
--["No sense being pessimistic, it wouldn't work anyway"]-- Opinions==mine


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss