I don’t know if it’s possible or not. Syscall-emulation mode is not fully “complete”. That is, not all syscalls have been implemented and, therefore, not all workloads will work with Syscall-emulation mode.
Beyond this I’d think about what kind of information you want from gem5. Syscall-emulation mode runs faster than Full-System mode but you can lose a lot of information. It’s not a cost-free speedup. If in doubt, use Full-System simulation. Given machine learning is computationally intensive, I understand the necessity for faster simulation (gem5 can easily run for 100k seconds to simulate just 1s of a workload run, dependent on what kind of system you're simulating). I would recommend figuring out what code you specially want to simulate and consider techniques such as checkpointing and KVM fast forwarding (if you’re host machine supports this) to focus simulation on smaller regions-of-interest. > On Apr 22, 2023, at 10:53 AM, saras nanda via gem5-users > <gem5-users@gem5.org> wrote: > > Hi Ayaz, thank you for the reply... is there any possibility to run it in > system emulation mode? > > regards > saras > > On Fri, Apr 21, 2023 at 12:43 PM Ayaz Akram <yazak...@ucdavis.edu > <mailto:yazak...@ucdavis.edu>> wrote: > Hi Saras, > > If you have a disk image with a machine learning model and relevant libraries > installed, you should be able to use it with gem5. I would suggest looking at > the examples of full-system simulation with gem5. Secondly, you might want to > look at how to modify disk images with QEMU and test your workload on QEMU. > If the model works on QEMU, I think theoretically it should work on gem5 as > well. > > -Ayaz > > On Thu, Apr 20, 2023 at 2:53 PM saras nanda via gem5-users > <gem5-users@gem5.org <mailto:gem5-users@gem5.org>> wrote: > Hi , > > > I am new to gem5. I would like to know if I can run a machine-learning model > on gem5. Would gem5 accept external libraries like TensorFlow, PyTorch..etc. > please provide me with some first steps to approach this problem of running a > machine learning model on gem5. > > regards > saras > > _______________________________________________ > gem5-users mailing list -- gem5-users@gem5.org <mailto:gem5-users@gem5.org> > To unsubscribe send an email to gem5-users-le...@gem5.org > <mailto:gem5-users-le...@gem5.org> > _______________________________________________ > gem5-users mailing list -- gem5-users@gem5.org > To unsubscribe send an email to gem5-users-le...@gem5.org
_______________________________________________ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org
