I faced a similar issue which caused simpoint based checkpoint generation to be inaccurate. I was able to get around it by simply commenting out the line that switches to simulation mode when the cpu context is halted. This fix worked for single-core configurations and I was able to get accurate fast forwarding for all of SPEC 2006. I haven't tested this fix on multi-core configurations yet, but I suspect that's where it might cause some kind of deadlock. Any insights on why the CPU context is halted would be very useful. Thanks!
On Mon, Apr 29, 2013 at 3:41 PM, Marisabel Guevara <[email protected]> wrote: > Hi all, > > I have been using the -fast-fwd-insns option in single-core simulations, > and I am getting a warning about switching to early simulation mode. > Looking through ptl-qemu.cpp, I see this happens when the CPU context is > stopped. Can anyone provide any insight into why the CPU context would > halt, and why it makes sense to switch to simulation mode at this point? I > plan on digging into why the context is stalled, or why it appears to be > stop. Any insight is much appreciated. > > > Best, > > -marisabel > > _______________________________________________ > http://www.marss86.org > Marss86-Devel mailing list > [email protected] > https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel > > -- Thanks, Akanksha Jain PhD Student Department of Computer Science University of Texas at Austin
_______________________________________________ http://www.marss86.org Marss86-Devel mailing list [email protected] https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
