Hello,
When I use 'm5 dumpstats 0' in my rcS scripts (e.g., scripts in
configs/boot/) I can easily match their execution with stats printing ('--
Begin ... --' and '--- End ---' stuff) in the stats.txt in FS mode.
It seems that you are using the equivament version from m5op.h. To my
understanding they are essentially the same, then you should also easily
match the m5_* calls with stats printing.
(When I observe strange thinks like you are observing, I try to run a
simple/dumb code/script simulation in gem5. If the problem persists, I run
again with a stable gem5 release - you may be observing a bug)
PS: in some conditions the linux distro may repawn the task in FS
simulation, for example if you don't put a 'm5 exit' or if the process is
killed.
Hope it helps,
--
Fernando A. Endo, Post-doc
INRIA Rennes-Bretagne Atlantique
France
2015-12-07 10:34 GMT+01:00 Tod <[email protected]>:
> Actually, I am using it like this "m5_dump_stats(0,0)", and I think, this
> should not repeat dumping.
>
> Also I am using the same in SE mode and there are only a couple of
> printings.
>
>
> Basically, I have made two points in both of my previous attempts in this
> thread, before asking the question.
>
> Could anyone please guide me on those as well? Whether my understanding is
> wrong?
>
>
> I would again like to say here:
>
> 1. In SE mode, it's much easier to associate each stats printing to
> its corresponding m5op.
> 2. However, in FS mode where there are multiple
> <http://comments.gmane.org/gmane.comp.emulators.m5.users/18824> (tens
> or even hundreds of) stats printings taking place in the 'stats.txt' file.
> Even
> if user does nothing, right?
>
> How can we identify the following:
>
>
> Which of the stats printing corresponds to what?
>
> OR
>
> At least, which of the stats printings are the sequel of m5ops invoked by
> a user?
>
>
>
>
> --------- 원본 메일 ---------
>
> *보낸사람*: Ayaz Akram <[email protected]>
> *받는사람* : gem5 users mailing list <[email protected]>
> *날짜*: 2015년 12월 07일 월요일, 08시 15분 27초 +0900
>
> *제목*: Re: [gem5-users] How to associate stats printings to m5ops inFSmode?
> Probably you are seeing too much dumped stats instances because m5ops like
> dumpresetstats repeat dumping every fixed number of ticks e.g. this is what
> documentation says about dumpstats :
> "dumpstats [*delay* [*period*]]: Save simulation statistics to a file in
> *delay* ticks; repeat this every *period* ticks."
> If you are only interested in ROI stats, you can call m5ops in your
> benchmark (before and after ROI) and then change the implementation
> of that m5op (or create a new m5op/pseudo ins ) in gem5 such that it does
> not dump stats instead only change the value of stats. And finally once
> your simulation ends, you will be able to see dumped stats anyway. I am not
> sure if you are looking for something like this or you are confused by
> something else. Anyway I answered based on my understanding !
>
> On Sat, Dec 5, 2015 at 2:37 AM, Tod <[email protected]
> <http:///hanmail/mail/[email protected]>>
> wrote:
>
>> Thanks Fernando for your reply.
>>
>>
>> I am using Gem5 FS mode with ARM detailed.
>>
>> I have been trying m5_dumpreset_stats and m5_dump_stats to know the
>> execution time in ticks for my ROI.
>>
>> The stats file produced has tens or even hundreds of stats printings[1],
>> which is justified[2] in FS mode.
>>
>>
>> Whereas, in SE mode this is not the case, where a couple of stats
>> printings[3] happen which can easily be associated to the inserted
>> m5op[4].
>>
>>
>> So, is there a way a user in FS mode can associate a stats printing to
>> the m5op used? To know that, Aha! this one is about my ROI in FS mode.
>>
>>
>> [1] stats printing refers to:
>>
>>
>> ---------- Begin Simulation Statistics ----------
>>
>>
>> ............................................................................
>>
>> - - - - - - - - - - - -- - - -- - - - - -- - - -- - - - - -
>>
>>
>> ............................................................................
>>
>> ---------- End Simulation Statistics -------------
>>
>>
>> [2] Please see the answer (Where Andreas has explained the fact) to this
>> question:
>>
>> http://comments.gmane.org/gmane.comp.emulators.m5.users/18824
>>
>>
>> [3] Please see the answer (Where Joel has explained the fact) to this
>> question:
>>
>> https://www.mail-archive.com/[email protected]/msg08083.html
>>
>>
>> [4] Anything like m5_dumpreset_stats or m5_dump_stats etc.
>>
>>
>>
>> --------- 원본 메일 ---------
>>
>> *보낸사람*: Fernando Endo <[email protected]
>> <http:///hanmail/mail/[email protected]>>
>> *받는사람* : gem5 users mailing list <[email protected]
>> <http:///hanmail/mail/[email protected]>>
>> *날짜*: 2015년 12월 05일 토요일, 01시 42분 01초 +0900
>> *제목*: Re: [gem5-users] How to associate stats printings to m5ops in
>> FSmode?
>>
>> Hello,
>>
>> Please, could you give us more information or/and an example of what
>> you're observing?
>>
>> Regards,
>>
>> --
>> Fernando A. Endo, Post-doc
>>
>> INRIA Rennes-Bretagne Atlantique
>> France
>>
>>
>> 2015-12-02 6:19 GMT+01:00 Tod <[email protected]
>> <http:///hanmail/mail/[email protected]>>:
>>
>>> Hi everyone,
>>>
>>>
>>> In SE mode, it's much easier to associate each stats printing to its
>>> corresponding m5op.
>>>
>>> However, in FS mode where there are multiple
>>> <http://comments.gmane.org/gmane.comp.emulators.m5.users/18824> (tens
>>> or even hundreds of) stats printings take place in the same 'stats.txt'
>>> file; How can we identify the following:
>>>
>>> Which of the stats printing corresponds to what?
>>>
>>> OR
>>>
>>> At least, which of the stats printings are the sequel of m5ops
>>> invoked by a user?
>>>
>>> _______________________________________________
>>> gem5-users mailing list
>>> [email protected]
>>> <http:///hanmail/mail/[email protected]>
>>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>>
>>
>> _______________________________________________
>> gem5-users mailing list
>> [email protected]
>> <http:///hanmail/mail/[email protected]>
>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>
>>
>>
>>
>> _______________________________________________
>> gem5-users mailing list
>> [email protected]
>> <http:///hanmail/mail/[email protected]>
>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>
>
> _______________________________________________
> gem5-users mailing list
> [email protected]
> <http:///hanmail/mail/[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
