[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [avr-gcc-list] AVR atmega128 GCC 4.3.3 Testresults comparision
From: |
Paulo Marques |
Subject: |
Re: [avr-gcc-list] AVR atmega128 GCC 4.3.3 Testresults comparision |
Date: |
Thu, 05 Nov 2009 15:24:28 +0000 |
User-agent: |
Thunderbird 2.0.0.21 (X11/20090310) |
Boyapati, Anitha wrote:
>
>> Try the files and simulator craeted as part of WINAVR
>>
>> http://winavr.cvs.sourceforge.net/viewvc/winavr/avrtest/
>>
>> Edit some paths to match your filespace.
>
> Thanks Andy. I think avrtest worked much better. I managed to get 42K passes!
BTW, I recently did an update on avrtest to include "cycle counters".
The wrappers to use them are in dejagnuboards/exit.c.
It allows you to write code like:
start_perf_counter(0);
...
start_perf_counter(1);
...do stuff...
stop_perf_counter(1);
...
start_perf_counter(2);
...do more stuff...
start_perf_counter(2);
...
stop_perf_counter(0);
The output would be something like:
counter 1: 1453 cycles
counter 2: 532 cycles
counter 0: 3274 cycles
So in a single run you can profile several functions and you have up to
256 counters to chose from.
Starting/stopping the counters consists of writing a single byte to a
port. This isn't exactly free, but if you use mostly counter 0, since
there is a __zero_reg__ always available, it will take a single cycle.
This cycle can be subtracted later to find the exact number of cycles
between start/stop.
Happy benchmarking,
--
Paulo Marques
Software Development Department - Grupo PIE, S.A.
Phone: +351 252 290600, Fax: +351 252 290601
Web: www.grupopie.com
"As far as we know, our computer has never had an undetected error."
Weisert
- Re: [avr-gcc-list] AVR atmega128 GCC 4.3.3 Testresults comparision,
Paulo Marques <=