Any change to the protocol .sm files requires the reviewer to know/understand/read the protocol. Alternatively, the patch can be downloaded + applied to tested with checkpoint creation.
As you said, checkpoints can be created with MOESI_hammer, and then restored with any other protocol, thus from the perspective of a one-time checkpoint creation/restore perspective that makes sense. That logic does not apply to flushing support in general, or when the user wishes to generate checkpoints for a protocol at regular intervals correct? I have received emails/requests/questions about this and another ruby patch where I wrote functionality to correctly account for cache hits/misses in the ruby stats, both of which have been on the review board for months without comments/suggestions. They are not some custom implementation/feature that only a handful of users may find useful. One of the points of a review board is to have the community post code/functionality that would be useful to other users (even if the code may not be 100% correct / or conform to everyone's liking), but that's why there is a review process, right? On Tue, Apr 16, 2013 at 5:46 PM, Nilay Vaish <[email protected]> wrote: > * Some body would need to read the protocol to figure out where or not the > changes are correct. > > * As of now, the only use case for flushing caches is checkpoint creation. > MOESI_hammer already supports flushing of caches and can be used to create > checkpoints. Those checkpoints can be used with any other protocol. So I > don't think any other protocol needs to support flushing. > > -- > Nilay > > > On Sun, 14 Apr 2013, Ali Saidi wrote: > > Brad, Nilay, and others? >> >> Is there a reason we haven't committed this? >> >> Thanks, >> Ali >> >> >> On Apr 13, 2013, at 8:41 AM, Malek Musleh <[email protected]> wrote: >> >> Atish, >>> >>> This has not been committed yet. It has not been reviewed/commented on by >>> any of the gem5-developers. As you can see, this was posted several >>> months ago, so I am not sure if they want this commited. Feel free to >>> download + apply it separately. >>> >>> Malek >>> >>> >>> >>> >>> On Fri, Apr 12, 2013 at 4:25 PM, Atish Patra <[email protected]> wrote: >>> >>> This is an automatically generated e-mail. To reply, visit: >>>> http://reviews.gem5.org/r/**1543/ <http://reviews.gem5.org/r/1543/> >>>> >>>> Ship it! >>>> >>>> Hi Malek Musleh, >>>> I am also working adding check point support on MOESI CMP directory >>>> protocol. >>>> Kindly let me know if you have committed these changes to gem5 >>>> development directory. >>>> Thanks in advance. >>>> Regards, >>>> Atish >>>> >>>> >>>> - Atish >>>> >>>> On December 2nd, 2012, 5:33 p.m., Malek Musleh wrote: >>>> Review request for Default. >>>> By Malek Musleh. >>>> >>>> *Updated Dec. 2, 2012, 5:33 p.m.* >>>> Description >>>> >>>> Changeset 9359:760ce17c1c4a >>>> --------------------------- >>>> Ruby: Add Flushing Support to MOESI CMP directory Protocol >>>> >>>> This patch adds some flushing functionality to the MOESI CMP directory >>>> Protocol. It's implementation is based off of the Flushing Implementation >>>> done for the MOESI Hammer Protocol. >>>> >>>> Testing: >>>> Currently, I only tested creating and restoring from a checkpoint made >>>> for the CG Benchmark from the NAS Suite. >>>> >>>> Diffs >>>> >>>> - src/mem/protocol/MOESI_CMP_**directory-L1cache.sm >>>> (**94383c5124d2f0cfe4af5ddeb09168**8a3a6c86b1) >>>> - src/mem/protocol/MOESI_CMP_**directory-L2cache.sm >>>> (**94383c5124d2f0cfe4af5ddeb09168**8a3a6c86b1) >>>> - src/mem/protocol/MOESI_CMP_**directory-dir.sm >>>> (**94383c5124d2f0cfe4af5ddeb09168**8a3a6c86b1) >>>> - src/mem/protocol/MOESI_CMP_**directory-msg.sm >>>> (**94383c5124d2f0cfe4af5ddeb09168**8a3a6c86b1) >>>> >>>> View Diff >>>> <http://reviews.gem5.org/r/**1543/diff/<http://reviews.gem5.org/r/1543/diff/> >>>> > >>>> >>>> ______________________________**_________________ >>> gem5-dev mailing list >>> [email protected] >>> http://m5sim.org/mailman/**listinfo/gem5-dev<http://m5sim.org/mailman/listinfo/gem5-dev> >>> >>> >> ______________________________**_________________ >> gem5-dev mailing list >> [email protected] >> http://m5sim.org/mailman/**listinfo/gem5-dev<http://m5sim.org/mailman/listinfo/gem5-dev> >> >> ______________________________**_________________ > gem5-dev mailing list > [email protected] > http://m5sim.org/mailman/**listinfo/gem5-dev<http://m5sim.org/mailman/listinfo/gem5-dev> > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
