Hello, I wanna run sha3 algorithm (hashing function) on gem5 for million instructions. The algorithm runs as follow: sha3sum 256 file path.
I added these parsers on se.py parser.add_option("-a", type="int") parser.add_option("-b", type="string") I run this command build/X86/gem5.opt configs/example/se.py -I 1000000 -c /home/abdkhail/project2/gem5/SHA3IUF/sha3sum -a 256 -b myfile --cpu-type=DerivO3CPU --caches The algorithm runs only 100 thousand instructions and then exits even if I passed the large file to the algorithm. Exiting @ tick 188440500 because exiting with last active thread context Simulated exit code not 0! Exit code is 1 Any help, please. Best Regards ________________________________ From: gem5-users <gem5-users-boun...@gem5.org> on behalf of Ciro Santilli <ciro.santi...@gmail.com> Sent: Monday, October 21, 2019 10:21 PM To: Francisco Carlos <juninho.u...@hotmail.com> Cc: gem5 users mailing list <gem5-users@gem5.org> Subject: Re: [gem5-users] stats seem to be wrong in multicore simulation using SE mode Maybe num_idle_cycles? But not sure, I would try to check if the values for that match and read source. On Mon, Oct 21, 2019 at 10:58 PM Francisco Carlos <juninho.u...@hotmail.com> wrote: > > Thank for your quick reply, Ciro. > > Yes, i call pthread_join() > > I am using risc-v isa and i saw, using debug flags, that there is a ecall > instruction where the cpu sleeps. > > So, the difference in cycles is the cycles that the cpus are sleeping, right? > > Is there a easy way to compute the sleep cycles to each cpu? Or is just > sim_ticks - cpu#.numCycles? > > Thanks in advance. > > > Obter o Outlook para iOS > ________________________________ > De: Ciro Santilli <ciro.santi...@gmail.com> > Enviado: Monday, October 21, 2019 6:50:37 PM > Para: gem5 users mailing list <gem5-users@gem5.org>; juninho.u...@hotmail.com > <juninho.u...@hotmail.com> > Assunto: Re: [gem5-users] stats seem to be wrong in multicore simulation > using SE mode > > Are the threads waiting for one another e.g. with pthread_join? > > This leads to futex syscalls WAIT, which I think make the CPU stop > ticking until WAKE, have a look at futexFunc. > > Also try tracing with --debug-flags ExecAll,SyscallBase you should be > able to see CPUs stop tracing after they go to sleep on syscalls. > > On Mon, Oct 21, 2019 at 7:20 PM Francisco Carlos > <juninho.u...@hotmail.com> wrote: > > > > Hi all, > > > > I am running a simulation using Pthreads lib in a four-core processor using > > SE mode in gem5. I separate the snippet of the stats file to show my > > problem, as can be seen below: > > > > > > sim_ticks 723694000 > > # Number of ticks simulated > > system.cpu0.numCycles 186781 > > # number of cpu cycles simulated > > system.cpu1.numCycles 636089 > > # number of cpu cycles simulated > > system.cpu2.numCycles 635944 > > # number of cpu cycles simulated > > system.cpu3.numCycles 650838 > > # number of cpu cycles simulated > > > > > > I am using 1 GHz clock frequency, therefore, the sim_ticks correspond to > > 723.694 cycles to execute the application. However, none of the processors > > took this number of cycles. > > > > Anyone know why is this happening? > > > > Thanks in advance. > > > > ---------------------------------------------------------------------------------------------------------------------------------------------- > > Francisco Carlos Silva Junior > > Phd student at University of Brasilia > > > > _______________________________________________ > > gem5-users mailing list > > gem5-users@gem5.org > > http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users _______________________________________________ gem5-users mailing list gem5-users@gem5.org http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________ gem5-users mailing list gem5-users@gem5.org http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users