See configs/common/Simpoint.py. Throwing a few print statements in there may help elucidate what's really going on. Just because the simpoint value changes in the config.ini file doesn't mean it's affecting anything.
Steve On Wed, Jul 1, 2009 at 9:24 AM, William George Beazley Jr < [email protected]> wrote: > Folks, > > here is a more detailed report on what makes my think simpoint is > functioning at least part of the time: > > I ran it three times, the final time with the simpoint being zero still > no change. Below the import of the configuration file I have the diffs > of the config.ini and m5stats.txt I presume there should be some change > in m5stats. > > MyBench1000M.py has a bench mark > > #164.gzip > gzip = LiveProcess() > gzip.executable = spec_2000_path + "/gzip/gzip00.peak.ev6" > data= spec_2000_path + "/gzip/input.graphic" > gzip.cmd = [gzip.executable] + [data] + ['60'] > gzip.output = 'input.graphic.out' > > #gzip.simpoint= 100*100E6 > #gzip.simpoint= 663*100E6 > gzip.simpoint= 0 > > [willi...@trout w_cpu2000]$ diff 1000_gzip/config.ini > 1000_gzip.663/config.ini > 133c133 > < simpoint=0 > --- > > simpoint=66300000000 > [willi...@trout w_cpu2000]$ > > > [willi...@trout w_cpu2000]$ diff 1000_gzip/m5stats.txt > 1000_gzip.663/m5stats.txt > 3,6c3,6 > < host_inst_rate 1045038 > # Simulator instruction rate (inst/s) > < host_mem_usage 4333128 > # Number of bytes of host memory used > < host_seconds 95.69 > # Real time elapsed on the host > < host_tick_rate 2304494659 > # Simulator tick rate (ticks/s) > --- > > host_inst_rate 1027809 > # Simulator instruction rate (inst/s) > > host_mem_usage 4333124 > # Number of bytes of host memory used > > host_seconds 97.29 > # Real time elapsed on the host > > host_tick_rate 2266501385 > # Simulator tick rate (ticks/s) > [willi...@trout w_cpu2000]$ env | grep 2000 > > > > > Thoughts? > > > > William George Beazley Jr wrote: > > > >Was this is ever resolved? I am having a similar problem. > > > >Further, I am trying to look for the code that runs simpoint; where is > >it implemented? > > > >I find this: > >[willi...@kani src]$ grep simpoint */* > >sim/Process.py: simpoint = Param.UInt64(0, 'simulation point at > >which to start simulation') > > > >I don't find it used again by name, however, how the c++ talks to > >Python is magic to me, so I suspect that it may get to it another way > >than by name. > > > >Thanks, > >Will > > > > > > > > > >Pavan Kumar wrote: > > > > > >>i am using cmp.py and Mybench.py file from this website, > >>http://www.m5sim.org/wiki/index.php/SPEC2006_benchmarks > >> > >>In Mybench.py, for every benchmark i added simpoint > >>for gcc > >>gcc.simpoint= 5*100E6 > >> > >> > >>command line i am using is: > >>/home/pavanj/m5-2.0b6/build/ALPHA_SE/m5.opt -d > >>/home/pavanj/m5-2.0b6/configs/cpu2006-linux-alpha-input/gcc > >>/home/pavanj/m5-2.0b6/configs/cpu2006-linux-alpha-input/cmp.py -t > >--caches > >>--l2cache --benchmark gcc > >> > >>On Thu, May 14, 2009 at 9:45 AM, Korey Sewell <[email protected]> wrote: > >> > >>> So to Lisa's point, what are the command lines that you are using? and > >>> what config file (se.py?)?? > >>> > >>> -- > >>> =========== > >>> Korey L Sewell > >>> PhD Candidate > >>> Computer Science & Engineering > >>> University of Michigan > >>> _______________________________________________ > >>> 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 > > > >----------------------------------- > >Will Beazley|Sys. Software Analyst > >409.880.7847|[email protected] > >_______________________________________________ > >m5-users mailing list > >[email protected] > >http://m5sim.org/cgi-bin/mailman/listinfo/m5-users > > > > ----------------------------------- > Will Beazley|Sys. Software Analyst > 409.880.7847|[email protected] > _______________________________________________ > 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
