I noticed that you tried to run 'create_checkpoint' script on '.img' (raw
image) file. Checkpoints are only supported in '.qcow2' images.
You can convert from .img to .qcow using 'qemu-img' which will be in 'qemu/'
folder.

Once the checkpoints are created, you can confirm its available by running
following command:
   $ qemu-img info your_qcow2_file

It will list the available checkpoints/snapshots in the qcow2 image.
These scripts use old parameters to create a checkpoint which is slow. You
need to use 'cache=unsafe' option to create/load checkpoints faster.
I have uploaded new scripts on github that reflects this change.

- Avadh

On Fri, Aug 26, 2011 at 4:50 PM, sparsh1 mittal1 <[email protected]>wrote:

> Hello
> While running run_bench script I get the following error. I will be
> grateful for the help.
>
> -bash-3.2$ python run_bench.py
> Execution command:
> Generic simulation command: simconfig -stats %s.stats -run -logfile %s.log
> Chekcpoints to run: ['blackscholes']
> All files will be saved in: ./OPrunbench/
> Running thread with img: /home/sparsh/cdrive/multicore/marss/parsecROI.img
> Starting Checkpoint: blackscholes
> All Threads are started
> initing the ooomachine
> char device redirected to /dev/pts/5
> char device redirected to /dev/pts/6
> QEMU 0.12.3 monitor - type 'help' for more information
> (qemu) loadvm blackscholes
> Snapshots not supported on device 'ide0-hd0'
> (qemu) simconfig -stats ./OPrunbench/blackscholes.stats -run -logfile
> ./OPrunbench/blackscholes.log
> simulation options received:-stats ./OPrunbench/blackscholes.stats -run
> -logfile ./OPrunbench/blackscholes.log
> (qemu)Writing to stdout completed
> Writing to output file completed
> WoooHooo... Finished Running all Simulations
>
> *Seeing the log file (part of it copied below) and Avadh's comment on rip,
> I think may be checkpoint is not proper :*
>
> Completed             0 cycles,             0 commits:         0
> Hz,         0 insns/sec: rip 00000000000f1eaa
> Completed        127000 cycles,         37959 commits:    630980 Hz,
> 188593 insns/sec: rip 00000000000f1eaa
>
> *However, create_checkpoint ran fine and its output was as follows:*
>
> Execution command:
> /home/sparsh/cdrive/multicore/marss/marss/qemu/qemu-system-x86_64 -m 2048
> -serial pty -vnc :20 -cpu core2duo -hda
> /home/sparsh/cdrive/multicore/marss/parsecROI.img
> Number of Chekcpoints to create: 1
> Starting QEMU for checkpoint: blackscholes
> initing the ooomachine
> char device redirected to /dev/pts/5
> ΓΏ
> Ubuntu 9.10 ubuntu ttyS0
>
> ubuntu login: root
> Password:
> Last login: Fri Aug 26 18:11:28 EDT 2011 on tty1
> Linux ubuntu 2.6.31.4qemu #3 SMP Mon Oct 19 16:41:41 EDT 2009 x86_64
>
> To access official Ubuntu documentation, please visit:
> http://help.ubuntu.com/
> cd parsec-2.1 ; . env.sh
>         export PARSEC_CPU_NUM=`grep processor /proc/cpuinfo | wc -l`; echo
> $PARSEC_CPU_NUM
>
> export CHECKPOINT_NAME="blackscholes"
>
> parsecmgmt -a run -c gcc-hooks -x roi -n 2 -p blackscholes
> root@ubuntu:~#
> ptlcall_mmio_init : Registered PTLcall MMIO page at physaddr 38654701568
> descriptor 120 io_mem_index 15
> ptlcall_mmio_write: calltype 3 at rip 140009376165944 (inside_ptlsim = 0 )
> PTLCALL type PTLCALL_CHECKPOINT
> Pending call type: 3
> MARSSx86::Creating checkpoint blackscholes
> MARSSx86::Checkpoint blackscholes created
> MARSSx86::Shutdown requested
>
> Thanks and Regards
> Sparsh Mittal
>
>
>
> _______________________________________________
> http://www.marss86.org
> Marss86-Devel mailing list
> [email protected]
> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>
>
_______________________________________________
http://www.marss86.org
Marss86-Devel mailing list
[email protected]
https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel

Reply via email to