Hi, This is not a benchmark problem. We have run a simple test yesterday, with this rcS:
#!/bin/sh /sbin/m5 resetstats echo "foo" sleep 1 echo "bar" /sbin/m5 exit "foo" is displayed, and after few minutes, gem5 stopped its execution *without* any error in the terminal. We have tried with the stable (Sept. 2015) and the current version. None of them is able to echo "bar". Note that the stable version contains modifications from us, whereas the current version is unmodified. The whole command line is: ./build/ARM/gem5.opt configs/example/fs.py -n 4 --cpu-type atomic --kernel ~/Big_LITTLE/arch/binaries/linux-arm64-gem5/vmlinux --machine-type VExpress_EMM64 --mem-type LPDDR3_1600_x32 --mem-size 256MB --disk-image ~/Big_LITTLE/arch/disks/linaro-minimal-aarch64.img --dtb-filename ~/Big_LITTLE/arch/binaries/linux-arm64-gem5/arch/arm64/boot/dts/aarch64_gem5_server.dtb --script ~/Big_LITTLE/gem5_current/gem5/configs/boot/test.rcS The Linux kernel is linux-arm64-gem5 [1], compiled with: make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- gem5_defconfig make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- -j 8 vmlinux With the same command line on 32-bit platform/kernel, everything looks good (with the stable modified and current version). [1] https://github.com/gem5/linux-arm64-gem5 On 07/01/2016 15:06, Louisa Bessad wrote: > Hi, > > Unfortunately none of these options fit to my problem. > > The first solution is not possible because we have other simulations > that last several days and gem5 has never exited before the end of the > simulation. > For the second one nothing appears on the console before the exit of gem5. > > Thanks > > Le 07/01/2016 14:44, Debiprasanna Sahoo a écrit : >> Hi, >> >> The command shows that you have not specified the -I option. That >> means you have not specified how many instruction to simulate >> before exiting. The probable reason (as of me) when such behavior >> occurs can be : 1. The benchmarks didn't complete its execution >> before gem5 terminates and hence it is not able to execute the exit >> echo in the script 2. There might be some error in the execution in >> gem5 which might have had terminated gem5. If there would be some >> error it must show on the console before exiting. >> >> You have to figure out which appropriately fits your scene. >> >> Thanks, Debiprasanna Sahoo Research Scholar IIT Bhubaneswar >> >> On Thu, Jan 7, 2016 at 7:04 PM, Louisa Bessad >> <[email protected] <mailto:[email protected]>> wrote: >> >> I execute: >> >> ./build/ARM/gem5.opt config/examples/ps.py -n 1 --cpu-type=atomic >> --kernel vmlinux64 --machine-type VExpress_EMM64 --mem-type >> LPDDR3_1600_x32 --mem-size 256MB --disk-image linaro_parsec >> --dtb-filename aarch64.dtb -b blackscholes >> >> For the kernel and the dtb file I use those available on gem5 git. >> I do the same for the disk image and I include the PARSEC compiled >> code for ARM64bits. >> >> Thanks >> >> Le 07/01/2016 14:26, Debiprasanna Sahoo a écrit : >>> Hi, >>> >>> Can you send the complete gem5 command, you executed? >>> >>> Thanks, Debiprasanna Sahoo Research Scholar IIT Bhubaneswar >>> >>> On Thu, Jan 7, 2016 at 6:52 PM, Louisa Bessad >>> <[email protected] <mailto:[email protected]> >>> <mailto:[email protected] <mailto:[email protected]>>> >>> wrote: >>> >>> Hi all, >>> >>> I am using gem5 to simulate an architecture ARM64bits and launch >>> the PARSEC benchmark. I place the compiled code of PARSEC in an >>> iso and I have made some rcS to launch each kernel of PARSEC, for >>> instance: >>> >>> #!/bin/sh cd /benchmarks/blackscholes /sbin/m5 resetstats echo >>> "Job launched" ./blackscholes "nb_thread" "in_file" "out_file" >>> echo "Job done" /sbin/m5 exit >>> >>> When I execute gem5, in the second terminal I see the launch of >>> the PARSEC kernel and the first echo. But there is no second echo >>> and then we have the exit. >>> >>> Do you have some ideas to explain why the last echo does not >>> appear whereas the exit is executed? >>> >>> Thanks in advance! >>> _______________________________________________ gem5-users >>> mailing list [email protected] <mailto:[email protected]> >> <mailto:[email protected] <mailto:[email protected]>> >>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users >>> >>> >>> >>> >>> _______________________________________________ gem5-users >>> mailing list [email protected] <mailto:[email protected]> >>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users >>> >> _______________________________________________ gem5-users mailing >> list [email protected] <mailto:[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 > -- +-------------------------------------------------------------+ | Pierre-Yves Péneau - PhD student | first.last at lirmm.fr | | LIRMM / CNRS - SYSMIC team | + 33 4 67 41 85 85 | | Building 4 Office H2.2 | http://walafc0.org | +-------------------------------------------------------------+
signature.asc
Description: OpenPGP digital signature
_______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
