On Sun, 17 May 2015, Amin Farmahini wrote:
Nilay,
I agree with Anthony. Even though you mention O3 does not support this as
of now, I think we should still have both opLat and issueLat.
Amin, neither you nor Tony has given any reason for having the issueLat
parameter. Therefore I would not revert this patch.
On a different matter, since I am not a frequent contributor to gem5, I
usually don't let myself make comments or review patches even though I go
over most non-ruby patches to educate myself. The thing that made me speak
up this time is that this patch was pushed without any "ship it" and it was
on review board for only 9 days. This does not seem right to me.
This is what happened about three weeks back. On 27th April, 2015, I
posted review request #2749 that modifies the ruby tester. Brad Beckmann
and Steve Reinhardt had negative opinions on the patch. I felt both Brad
and Steve have taken positions that are not right. I contemplated moving
away from gem5 since I did not want to get into a confrontation.
Therefore, on 28th April, I discarded all my review requests that I was
not going to discuss or commit. I also committed the issueLat patch and
few others which I felt were worth going in. This is the reason why the
issueLat patch was not kept longer on the reviewboard.
Before making up your mind on this incident, I would strongly suggest that
you read the entire conversation from the archives of the mailing list
and understand the two patches on ruby tester posted by AMD and me
respectively. Finally, you might ask why am I still here. I think I like
the tool too much to call it quits so easily.
--
Nilay
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev
