> On Oct. 24, 2012, 1:55 p.m., Brad Beckmann wrote: > > src/mem/protocol/MESI_CMP_directory-L1cache.sm, line 40 > > <http://reviews.gem5.org/r/1489/diff/1/?file=30553#file30553line40> > > > > Minor comment, but can you change this name to "enable_prefetch"
Done! > On Oct. 24, 2012, 1:55 p.m., Brad Beckmann wrote: > > src/mem/protocol/MESI_CMP_directory-L1cache.sm, line 273 > > <http://reviews.gem5.org/r/1489/diff/1/?file=30553#file30553line273> > > > > As I mentioned in my previous review, for historical reasons you may > > want to call this the "optionalQueue". Done! > On Oct. 24, 2012, 1:55 p.m., Brad Beckmann wrote: > > src/mem/protocol/MESI_CMP_directory-L1cache.sm, line 589 > > <http://reviews.gem5.org/r/1489/diff/1/?file=30553#file30553line589> > > > > Below the L1 cache, do you treat prefetches and non-prefetches the > > same? Have you looked into the implications of this? As of know the prefetch and demand requests are treated equally after the L1 cache. I have not done any experimentation on how the prefetch traffic affects demand traffic. While we can treat them differently, I think we should keep things simple to begin with. Over time, I expect protocols would change to adapt better. - Nilay ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/1489/#review3593 ----------------------------------------------------------- On Oct. 24, 2012, 12:40 p.m., Nilay Vaish wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.gem5.org/r/1489/ > ----------------------------------------------------------- > > (Updated Oct. 24, 2012, 12:40 p.m.) > > > Review request for Default. > > > Description > ------- > > Changeset 9315:fb4f9134eaec > --------------------------- > ruby: add support for prefetching to MESI protocol > > > Diffs > ----- > > configs/ruby/MESI_CMP_directory.py f634a34f2f0b > src/mem/protocol/MESI_CMP_directory-L1cache.sm f634a34f2f0b > src/mem/protocol/RubySlicc_Types.sm f634a34f2f0b > src/mem/slicc/ast/ObjDeclAST.py f634a34f2f0b > src/mem/slicc/symbols/StateMachine.py f634a34f2f0b > > Diff: http://reviews.gem5.org/r/1489/diff/ > > > Testing > ------- > > > Thanks, > > Nilay Vaish > > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
