Re: [gem5-dev] Review Request 3117: arch, x86: Delete packet in IntDevice::recvResponse

2015-09-28 Thread Joel Hestness
> On Sept. 28, 2015, 8:39 a.m., Andreas Hansson wrote: > > src/dev/x86/intdev.hh, line 150 > > > > > > This is fine, but what about just making it purely virtual? As proposed in the latest patch, this allows us to inherit

Re: [gem5-dev] Review Request 3087: ruby: Fix CacheMemory allocate leak

2015-09-28 Thread Joel Hestness
> On Sept. 2, 2015, 2:02 p.m., Jason Lowe-Power wrote: > > Is it not possible to delete it when the entry is set to NotPresent? It > > seems weird to delete it in the allocate function. > > Nilay Vaish wrote: > I think the problem here is with the coherence protocols maintaining two >

Re: [gem5-dev] Review Request 3087: ruby: Fix CacheMemory allocate leak

2015-09-28 Thread Nilay Vaish
> On Sept. 2, 2015, 2:02 p.m., Jason Lowe-Power wrote: > > Is it not possible to delete it when the entry is set to NotPresent? It > > seems weird to delete it in the allocate function. > > Nilay Vaish wrote: > I think the problem here is with the coherence protocols maintaining two >

Re: [gem5-dev] Review Request 3134: ruby: combine RubyPort and Sequencer and rename as FirstLevelController

2015-09-28 Thread Nilay Vaish
> On Sept. 24, 2015, 11:32 p.m., Brad Beckmann wrote: > > Do not check this in. I suggest you stop going in this direction. Any > > performance benefits provided by this patch do not justify the resulting > > duplicate code and downstream re-correlation and development efforts. > > There

Re: [gem5-dev] Review Request 3087: ruby: Fix CacheMemory allocate leak

2015-09-28 Thread Nilay Vaish
> On Sept. 2, 2015, 2:02 p.m., Jason Lowe-Power wrote: > > Is it not possible to delete it when the entry is set to NotPresent? It > > seems weird to delete it in the allocate function. > > Nilay Vaish wrote: > I think the problem here is with the coherence protocols maintaining two >

Re: [gem5-dev] Review Request 3115: ruby: RubyPort delete snoop requests

2015-09-28 Thread Andreas Hansson
--- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/3115/#review7298 --- Ship it! Could you update the comments in the code? Also, the patch

Re: [gem5-dev] Review Request 3134: ruby: combine RubyPort and Sequencer and rename as FirstLevelController

2015-09-28 Thread Nilay Vaish
> On Sept. 24, 2015, 11:32 p.m., Brad Beckmann wrote: > > Do not check this in. I suggest you stop going in this direction. Any > > performance benefits provided by this patch do not justify the resulting > > duplicate code and downstream re-correlation and development efforts. > > There

Re: [gem5-dev] Review Request 3117: arch, x86: Delete packet in IntDevice::recvResponse

2015-09-28 Thread Andreas Hansson
--- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/3117/#review7296 --- src/dev/x86/i82094aa.hh (line 108)

Re: [gem5-dev] Review Request 2087: mem: Add a wrapped DRAMSim2 memory controller

2015-09-28 Thread Andreas Hansson
> On Sept. 28, 2015, 7:42 a.m., Lang Ma wrote: > > Dear Andreas, > > > > I apply this patch to gem5-44ef5ed3aee0 and follw steps to get DRAMSim2 as > > part of gem5, but I face a problem as follws: > > > > se.py: error: option --mem-type: invalid choice: 'dramsim2' (choose from > >

Re: [gem5-dev] Review Request 3121: ruby: block size in RubySystem to be object specific

2015-09-28 Thread Brad Beckmann
> On Sept. 17, 2015, 5:43 p.m., Brad Beckmann wrote: > > Do *not* check in this patch. I know I've already saidy this, but I will > > repeat, do not make further changes to the sm files until we check in our > > GPU patches. > > > > Even after we check in our GPU patches, this patch must be