You have to look at the "test.py" file to see what benchmark the regression
is running and what input. The "ref" folder refers to a reference for the
regression test, not necessarily the "ref" inputs...
Consider you navigate to "tests/long/00.gzip".
In that folder you see a "test.py" file.
In that file, it says:
"m5.AddToPath('../configs/common')
from cpu2000 import gzip_log
workload = gzip_log(isa, opsys, 'smred')
root.system.cpu.workload = workload.makeLiveProcess()"
Thus, you know gzip was run with the small reduced ("smred") input set.
Then if you go to the folder
tests/long/00.gzip/ref/alpha/tru64/simple-atomic
you'll see the outputs for the atomic-simple-cpu model for that
regression...
On Fri, Mar 28, 2008 at 7:43 PM, Sujay Phadke <[EMAIL PROTECTED]>
wrote:
> Thanks.
>
> But if we run the smred (or other reduced benchmarks), there is no output
> directory in the /data/<banchmark>/smred
> directory to compare with. I can see the output directory only for the
> reference case.
>
> - Sujay
>
> ----- Original Message -----
> From: "Ali Saidi" <[EMAIL PROTECTED]>
> To: "M5 users mailing list" <[email protected]>
> Sent: Friday, March 28, 2008 5:41 PM
> Subject: Re: [m5-users] Running Spec2K in SE mode beta 5
>
>
> >
> > On Mar 28, 2008, at 4:44 PM, Sujay Phadke wrote:
> >> Hello all,
> >> I am trying to run the specint benchmarks in SE mode in beta 5,
> >> using m5.fast. I am facing errors with some of them as follows:
> >>
> >> I am running them on a machine with 2GB ram. I made the correction to
> >> the cpu2000.py script as suggested by Vilas in previous postings.
> >>
> >> 1. mcf -
> >>
> >> warn: Entering event queue @ 0. Starting simulation...
> >> warn: Increasing stack size by one page.
> >> fatal: Out of memory, please increase size of physical memory.
> >> @ cycle 45551000
> >> [new_page:build/ALPHA_SE/sim/system.cc, line 235]
> > It means exactly what it says. You have a physical memory of probably
> > 128MB, and the binary used all the available memory. Increase the size
> of
> > the PhysicalMemory Object.
> >
> >
> >>
> >> 2. parser -
> >> AttributeError: OptionParser instance has no __call__ method. I
> >> think it needs input thru STDIN but does not seem to work
> > You've managed to grab the wrong module. There is a paser benchmark,
> and
> > a parser python module. You've called the latter. Instead of from XXXX
> > import * lines you should use import XXXX and then call XXX.callable()
> >
> >>
> >> 3. amongst the specint benchmarks, perlbmk and bzip2 do not have a
> smred
> >> input set. So what input_set options do we have to pass on?
> > You can use any of the defined input sets. I believe the minnespec
> > benchmarks for bzip just used different parameters.
> >
> >>
> >> Also, how do we check if the benchmarks ran correctly or no?
> > By comparing the output to the reference output
> >
> >>
> >> Has anyone gotten all these to run correctly?
> >> thanks,
> >> Sujay
> >>
> >> _______________________________________________
> >> m5-users mailing list
> >> [email protected]
> >> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
> >
> > _______________________________________________
> > m5-users mailing list
> > [email protected]
> > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
> >
>
> _______________________________________________
> m5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
>
--
----------
Korey L Sewell
Graduate Student - PhD Candidate
Computer Science & Engineering
University of Michigan
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
