I have a somewhat similar question. The method Ali mentioned works well
when your ROI is large, but what if your ROI is short? For example the loop
body of a loop that is executed many times.
I believe using m5_dumpreset_stats is not doable in this case for two
reasons: (1) since the loop body is short, the overhead of
m5_dumpreset_stats is not negiligible. m5_dumpreset_stats is also executed
non-speculatively which imposes performance cost. (2) the loop is exectued
many times so it dumps stats every time the loop is executed.
I am interested in finding the execution time of the loop body (and not the
loop boundary checking, etc). So I was wondering if there is such a way to
do fine-grained profling within Gem5?

Thanks,
Amin

On Fri, Aug 17, 2012 at 12:31 PM, Anh Nguyen <[email protected]> wrote:

> Tony,
>
> It traces both kernel and user-space codes. I traced JVM before.
>
> Anh-
>
>
> On Fri, Aug 17, 2012 at 10:13 AM, Anthony Gutierrez <[email protected]>wrote:
>
>> I've used an x86_64 host with Ubuntu to build SystemTap modules and run
>> them on Android ARM, so it's doable. But, I don't think SystemTap is
>> relevant here if you're trying to do architectural simulation. SystemTap is
>> a profiling tool that uses kprobes; it doesn't really do userspace
>> profiling last time I checked.
>>
>> -Tony
>>
>> On Fri, Aug 17, 2012 at 1:05 PM, Amin Farmahini <[email protected]>wrote:
>>
>>> Ahn,
>>>
>>> I am not familiar with SystemTap, but I just took a look at its
>>> beginners guide and I believe you cannot use this tool for the kind of
>>> things Shervin would like to do. This is because "The host system must be
>>> the same architecture and running the same distribution of Linux as the
>>> target system in order for the built instrumentation module to work."
>>> My understanding could be wrong, though.
>>>
>>> Thanks,
>>> Amin
>>>
>>> On Fri, Aug 17, 2012 at 11:57 AM, Anh Nguyen <[email protected]>wrote:
>>>
>>>> Systemtap
>>>
>>>
>>>
>>> _______________________________________________
>>> gem5-users mailing list
>>> [email protected]
>>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>>
>>
>>
>> _______________________________________________
>> gem5-users mailing list
>> [email protected]
>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>
>
>
> _______________________________________________
> gem5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to