[ECOS] Execution time

[ECOS] Execution time

[Michele Portolan]

I would need to check the execution time of different implementations of 
the same function: is there a simple way to do it?
I lloked up "tm_basic.cxx", but I think my case is far simpler because I 
have no scheduler (for the moment).
Thanks anyone,

Michele

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

Re: [ECOS] Execution time

[Andrew Lunn]

On Wed, Sep 20, 2006 at 08:34:21AM +0200, Michele Portolan wrote:
> I would need to check the execution time of different implementations of 
> the same function: is there a simple way to do it?
> I lloked up "tm_basic.cxx", but I think my case is far simpler because I 
> have no scheduler (for the moment).

If you want to keep it simple use HAL_CLOCK_READ(). But you will have
to handle wrap around and scaling between HAL CLOCK ticks and real
ticks.

   Andrew

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

Re: [ECOS] Execution time

[Nick Garnett]

Michele Portolan <micheleportolan@yahoo.it> writes:

> I would need to check the execution time of different implementations
> of the same function: is there a simple way to do it?
> I lloked up "tm_basic.cxx", but I think my case is far simpler because
> I have no scheduler (for the moment).
> Thanks anyone,


tm_basic shows the standard way of doing this: execute the test
routine many times in a loop, time the whole thing, and calculate the
time per run: total_time/loop_count .

Timing a single excecution is prone to all sorts of interference from
interrupts, cache effects, pipeline effects, DRAM refresh
etc. Multiple runs helps to minimize the effect of these things.

-- 
Nick Garnett                                 eCos Kernel Architect
http://www.ecoscentric.com            The eCos and RedBoot experts


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

RE: [ECOS] Execution time

[Kalyani, Korubilli (IE10)]

Hi,

I want to know how to implement Application exception handler in ECOS.
Plz provide me the sufficient information regarding this as soon as
possible please.

 

Regards,

Kalyani

-----Original Message-----
From: ecos-discuss-owner@ecos.sourceware.org
[mailto:ecos-discuss-owner@ecos.sourceware.org] On Behalf Of Nick
Garnett
Sent: Wednesday, September 20, 2006 10:13 PM
To: Michele Portolan
Cc: ecos-discuss@ecos.sourceware.org
Subject: Re: [ECOS] Execution time

Michele Portolan <micheleportolan@yahoo.it> writes:

> I would need to check the execution time of different implementations
> of the same function: is there a simple way to do it?
> I lloked up "tm_basic.cxx", but I think my case is far simpler because
> I have no scheduler (for the moment).
> Thanks anyone,


tm_basic shows the standard way of doing this: execute the test
routine many times in a loop, time the whole thing, and calculate the
time per run: total_time/loop_count .

Timing a single excecution is prone to all sorts of interference from
interrupts, cache effects, pipeline effects, DRAM refresh
etc. Multiple runs helps to minimize the effect of these things.

-- 
Nick Garnett                                 eCos Kernel Architect
http://www.ecoscentric.com            The eCos and RedBoot experts


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


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