config: Use BaseSESystem in multi-processor regressions

Information
Submitter:	Andreas Hansson
Repository:	gem5
Branch:	default
Bugs:
Depends On:
Reviewers
Groups:	Default
People:

Description

Changeset 9773:840f67490b6a
---------------------------
config: Use BaseSESystem in multi-processor regressions

This patch adopts the BaseSESystem also in the o3-timing-mp,
simple-timing-mp and simple-atomic-mp regressions (only SPARC).

The default builder is used with existing caches sizes as they match
what was used in these scripts.

Due to the different order the ports are connected, the bus stats get
shuffled around. A separate patch bumps the port indices. Besides
this, all behaviour is exactly the same.

Testing Done

All regressions pass

Seems fine.

Andreas Hansson June 6, 2013, 11:10 p.m. (June 6, 2013, 11:10 p.m.)

Btw, it would be great if we could do the same with the .*-ruby configs, but I am not quite sure how to achieve that. Any ideas?

LGTM

LGTM

Same comments as before: (1) please append to rather than replace existing copyrights and (2) why not just merge this with the previous one.

Andreas Hansson June 11, 2013, 12:07 a.m. (June 11, 2013, 12:07 a.m.)

The replacing here was based on the fact that nothing of the original file is left. Do you still think the existing copyright should stay even though none of the contents did?

I'll merge the two if desired. The main reason I did not do so initially is that this one requires stats updates whereas the previous one does not.

Steve Reinhardt June 11, 2013, 12:30 a.m. (June 11, 2013, 12:30 a.m.)

Looking at the diffs, we still have two complete lines 'from m5.objects import *' and 'nb_cores = 4' left over from the original files (not to mention a common expression or two that is less than a full line).  I wouldn't say that's significant, but it seems to me that the ARM copyright has been added to files over equally minor amounts of code, so it gives a bad appearance for an ARM person in particular to not be just as pedantic when the shoe is on the other foot.  Also there's really no cost or harm to leaving the existing copyrights.

I'll leave the merging up to you; I don't see a gain in keeping them separate, and I think there is a slight loss in having closely related changes split across two patches, but it's not a big deal either way.

Description:

~		Changeset 9758:47d9b6068c50
	~	Changeset 9773:840f67490b6a

		config: Use BaseSESystem in multi-processor regressions

		This patch adopts the BaseSESystem also in the o3-timing-mp,
		simple-timing-mp and simple-atomic-mp regressions (only SPARC).

		The default builder is used with existing caches sizes as they match
		what was used in these scripts.

		Due to the different order the ports are connected, the bus stats get
		shuffled around. A separate patch bumps the port indices. Besides
		this, all behaviour is exactly the same.

Diff:

Revision 2 (+48 -126)

Show changes

	tests/configs/o3-timing-mp.py
	tests/configs/simple-atomic-mp.py
	tests/configs/simple-timing-mp.py

You have a pending review.

Review Board 2.0.15

This change has been discarded.

Screenshots

Files

Description:

Changeset 9758:47d9b6068c50

Changeset 9773:840f67490b6a

Diff:

Status: Discarded