I agree with Alex: the ISA description system was designed for Alpha, and
it remains the purest example of how it was intended to be used, so I think
there's some value in keeping it around for that.

To me, it should boil down to a cost/benefit consideration. I agree that
the benefits are not that great any more, but I think the costs are pretty
minimal too, and could be reduced further without abandoning Alpha
altogether. For example, seeing some of the things Brandon has run into
with his Process code restructuring & enhancement, I'm in favor of dropping
Tru64 support. Also, we certainly could get rid of many or all of the
long-running regression tests that aren't adding any value (which is not
even a problem that's unique to Alpha). But it seems overkill to me to get
rid of Alpha entirely, especially since it's such a clean ISA for the most
part.

Now if someone proposed getting rid of an ISA that would let us drop a
microarchitectural complexity like register windows, or branch delay slots,
or predication ;), (or to be fair, microcode!) then the cost side of the
equation might be different, but of course the benefits are more tangible
for those ISAs as well.

Steve


On Tue, Oct 25, 2016 at 9:46 AM Dutu, Alexandru <[email protected]>
wrote:

> Hi Andreas,
>
> Not sure I understand this decision/proposal. Certainly, there are some
> ISAs in the current repo which are not used or maintained by anyone. This
> might render them good candidates for removal. However, ALPHA is quite a
> big exception in my mind.
>
> Surely, very few people are using ALPHA in their research and no one
> maintains it, yet gem5 was developed with ALPHA in mind. There might still
> exist assumptions baked in the current code because of how ALPHA works. I
> addition, I would also argue that ALPHA has a good example of  ISA
> generation code. If not in research, it can be used in educating users
> about gem5.
>
> Therefore, keeping ALPHA around is reasonable to me, it can be helpful in
> refining the gem5 models and finding out bugs.
>
> Best regards,
> Alex
> -----Original Message-----
> From: gem5-dev [mailto:[email protected]] On Behalf Of Andreas
> Hansson
> Sent: Monday, October 24, 2016 1:50 PM
> To: [email protected]; [email protected]
> Subject: [gem5-dev] Removal of ALPHA from gem5
>
> Hi all,
>
> The day has come.
>
> As previously discussed, http://reviews.gem5.org/r/3684/ removes support
> for ALPHA from gem5. Ultimately this has to be a community decision. Thus,
> if anyone is actually relying on ALPHA for their work please let us know
> within the next two weeks.
>
> If there are any questions or concerns, please speak up. Depending on the
> feedback, I aim to push this patch on November 14th.
>
> Kind regards,
>
> Andreas
>
> IMPORTANT NOTICE: The contents of this email and any attachments are
> confidential and may also be privileged. If you are not the intended
> recipient, please notify the sender immediately and do not disclose the
> contents to any other person, use it for any purpose, or store or copy the
> information in any medium. Thank you.
> _______________________________________________
> gem5-dev mailing list
> [email protected]
> http://m5sim.org/mailman/listinfo/gem5-dev
> _______________________________________________
> gem5-dev mailing list
> [email protected]
> http://m5sim.org/mailman/listinfo/gem5-dev
>
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to