> On Feb. 24, 2015, 5:29 p.m., Steve Reinhardt wrote:
> > Just curious: what's the motivation for including this code with gem5 and 
> > not with SST?  I'm not necessarily opposed, just wondering whether you 
> > considered putting this code in the SST distribution instead.

Yes, both ways were considered.

gem5 "binds tighter" in many ways than SST which led to this approach being 
taken.  For instance, SST dynamically loads its Components (~SimObjects) 
whereas gem5 does not.  Keeping it in our tree is the better of the two options.


- Curtis


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviews.gem5.org/r/2667/#review5901
-----------------------------------------------------------


On Feb. 20, 2015, 6:47 p.m., Curtis Dunham wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviews.gem5.org/r/2667/
> -----------------------------------------------------------
> 
> (Updated Feb. 20, 2015, 6:47 p.m.)
> 
> 
> Review request for Default.
> 
> 
> Repository: gem5
> 
> 
> Description
> -------
> 
> Changeset 10708:344bc6b7203d
> ---------------------------
> ext: Add SST connector
> 
> This patch adds a connector that allows gem5 to be used as a component
> in SST (Structural Simulation Toolkit, sst-simulator.org). At a high
> level, this allows memory traffic to pass between the two simulators.
> SST Links are roughly analogous to gem5 Ports, although Links do not
> have a notion of master and slave. This distinction is important to
> gem5, so when connecting a gem5 CPU to an SST cache, an ExternalSlave
> must be used, and similarly when connecting the memory side of SST cache
> to a gem5 port (for memory <-> I/O), an ExternalMaster must be used.
> 
> These connectors handle the administrative aspects of gem5
> (initialization, simulation, shutdown) as well as translating SST's
> MemEvents into gem5 Packets and vice-versa.
> 
> 
> Diffs
> -----
> 
>   ext/sst/ExtMaster.hh PRE-CREATION 
>   ext/sst/ExtMaster.cc PRE-CREATION 
>   ext/sst/ExtSlave.hh PRE-CREATION 
>   ext/sst/ExtSlave.cc PRE-CREATION 
>   ext/sst/LICENSE PRE-CREATION 
>   ext/sst/Makefile PRE-CREATION 
>   ext/sst/README PRE-CREATION 
>   ext/sst/gem5.hh PRE-CREATION 
>   ext/sst/gem5.cc PRE-CREATION 
>   ext/sst/libgem5.cc PRE-CREATION 
> 
> Diff: http://reviews.gem5.org/r/2667/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Curtis Dunham
> 
>

_______________________________________________
gem5-dev mailing list
gem5-dev@gem5.org
http://m5sim.org/mailman/listinfo/gem5-dev

Reply via email to