On Mon, 29 Aug 2011, Jason Power wrote:
Thanks Nilay,
That got me a bit further. I seem to be able to run single threaded
applications without a problem with x86 and Ruby now, however, like you
said, more than one CPU doesn't seem to work.
If anyone is interested, I think I tracked it down to an issue with the
atomic operations, but I can't (or don't have time) trace it beyond that.
Jason, it is very well possible that there is some error some where. We
are always interested in knowing errors encountered by gem5's users. You
should certainly provide details on what you found.
In this particular case, my view is that only fixing this one issue that
you found would not fix the problem of O3 and Ruby working together. We
need to develop some mechanism to communicate information from the Ruby
memory system to the CPU. I think a good way to start working on this
would be to write some particular test cases for testing whether or not
the system is working correctly. Assuming that we are going to mostly work
with x86-based systems and would like a TSO consistency model, I think we
should implement tests that appear in x86 manuals from AMD and Intel.
Thanks
Nilay
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
