I know Geoff Blake had worked on some TM stuff on an older version of M5. It might be a good idea to check out some of the stuff he posted on his webpage: eecs.umich.edu/~blakeg
With regards to component testing, it might be worth configuring the SimpleCPU model as a way to test different things. I know It's probably not exactly 'component testing' in its purest sense, but it may not be too difficult to get that model to execute some instruction/memory trace that you want to use for small-scale testing. Once you figure out where instructions interface to memory you can edit the simulation at that point. I'm not sure about what you would have to do if you want to isolate parts of the memory system though. On Tue, Mar 30, 2010 at 11:24 PM, nathan binkert <[email protected]> wrote: > > > We are trying to implement a type of Hardware Transactional memory on M5 > > Simulator. We need to modify a lot of parts for that and would like to check > > small modules before running the entire program on a benchmark. Is there > > any way > > to do that? > > I'm not exactly sure what you're asking for. Component testing? If > so, there aren't really test harnesses for each individual M5 > components as they all pretty much depend on each other. For generic > simulator functions like the event queue, statistics, etc., there are > some tests in src/unittest > > Nate > _______________________________________________ > m5-users mailing list > [email protected] > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users -- - Korey _______________________________________________ m5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
