I adjusted the description to at least warn them what they're
getting into, but I guess you're right. If they're unusually smart,
capable, hardworking and crazy, why not let them give it a shot? :) I
also redid the description of the PLI interface project.
Gabe
Ali Saidi wrote:
I guess you're right... It should stay. Difficulty really isn't
something we need to filter, just grunt work.
Ali
On Mar 12, 2008, at 1:57 PM, Korey Sewell wrote:
It's a WIKI, so feel free to re-word the blurb if you like.
I dont think we should limit what we think people could do or not in
terms of their ability.
It's hard work, will take a good bit of thought, but if someone could
figure this out in at least
a limited capacity wouldnt that be a good thing? The solution might be
complex or it might
be a worked solution of having two separately compiled M5 binaries
interact with each other?
I honestly think it's a cool idea, but I guess I was thinking too big
so if all else fails and nobody really thinks it's a good
idea then just remove it ...
On Wed, Mar 12, 2008 at 7:00 PM, Gabe Black <[EMAIL PROTECTED]>
wrote:
It's probably really late to bring this up, but I know that
heterogeneous ISA support is going hard, painful, and will have to be
done very carefully not to make a big mess. There's probably quite a
bit
too much there for somebody to figure out in one summer, but if we
leave
it on the list we should try to reword the blurb to not make it sound
like it'll just take a few adjustments and it'll all work out. I'd say
it would be about half as hard as porting all of m5 into perl, for
instance.
Gabe
Korey Sewell wrote:
yea, fair enough... good points.
On Wed, Mar 12, 2008 at 2:18 AM, Ali Saidi <[EMAIL PROTECTED]> wrote:
Yea, but we don't want the GSoC organizers to think that we're
proposing a bunch of junk that we don't want to do. If I had time I
would actually enjoy doing a direct execution model, parallelizing
m5,
or making an in-order core model. I would certainly not enjoy
benchmarking and validation.
Ali
On Mar 12, 2008, at 2:06 AM, Korey Sewell wrote:
Also for new users or undergraduate studies, I think grunt work
wouldn't be that bad if you can port applications that are somewhat
interactive or provide "cool" output.
Didn't someone get "DOOM" working somehow with SimpleScalar back in
the day?
Stuff that has some kind of end "prize" can be motivational for
students just learning about computer architecture and such...
And in the end, if it ends helping to grow the M5 software
development
then that's great...
On Wed, Mar 12, 2008 at 1:29 AM, Ali Saidi <[EMAIL PROTECTED]> wrote:
I don't know that we should keep the Benchmarking and Verification
section. I removed the old one because who really gets excited
about
which can only be described as grunt work.
Ali
On Mar 12, 2008, at 1:27 AM, Korey Sewell wrote:
OK, I threw a couple of those in there as well. Looks like the
Project
List is filling out a bit... Only thing else I could think of
personally is add the PowerPC ISA port for M5 (hasn't someone
worked
on that or is someone currently doing that?).
On Wed, Mar 12, 2008 at 12:18 AM, nathan binkert
<[EMAIL PROTECTED]>
wrote:
Write some blurbs about these ideas.
On Tue, Mar 11, 2008 at 9:09 PM, Korey Sewell <[EMAIL PROTECTED]>
wrote:
I added "Interface to Power Analysis tool" to the Project Ideas
list...
I'm wondering if "Benchmarking & Verification on M5" or some
derivation of that would be relevant? The idea would be to
verify
that
workloads and benchmarks of interest work properly on M5. This
could
add to some phase of our regression testing. Not sure this would
be
great "fun" for something to do, but it could be useful...
What about Heterogenous ISA Systems?
Transactional Memory? (Geoff has this but is it ready to be put
into
M5? Maybe someone could help him create a EXTRAS module or
something?)
Processor Models (in addition to detailed-in-order): Graphics
Processor Model; Stream Processor Model; VLIW????
That's all I can think of... Let me know if that's fine to
throw
in there...
On Tue, Mar 11, 2008 at 11:20 PM, Ali Saidi <[EMAIL PROTECTED]>
wrote:
I've re-organized things a bit. I've completely changed the
introduction and i'm working on what is now "other". Please
work on
turning the bullet point descriptions into a paragraphs.
When you
do
please use the edit button next to the section you're
working on
(instead of editing the whole page), so that we don't step on
each
others changes.
Ali
On Mar 11, 2008, at 10:43 PM, nathan binkert wrote:
The intro text needs to be changed significantly. It looks
way
too
much like the mercurial text.
On Tue, Mar 11, 2008 at 6:57 PM, Korey Sewell
<[EMAIL PROTECTED]>
wrote:
I did some light work on the GSOC page setting it up more
similar to
the Mercurial style.
- I added a light M5 introduction
- Added Mentors List
- Starting "Getting Things Done" Section...
The Project Ideas list still needs to be done more in the
Mercurial
style and the Getting Things Done section probably should be
added as
well...
I hope what I did helps the cause a bit!...
On Tue, Mar 11, 2008 at 9:15 PM, nathan binkert
<[EMAIL PROTECTED]>
wrote:
For the "ideas" page, you could just list out device
models
that
would be interesting. I see Flash Memory device, but if you
guys
that
are familiar working with devices in M5 know of things off
the
top of
your head that seems like a list that may be able to be
populated
rather quickly.
If you have something in mind, go ahead and add it.
Also, how closely do we want to mirror Mercurial's style
for
Summer of
Code 2008? If we are using this as a guide, we would need a
separate
introduction section for M5 & GSOC (like Ali said) and also
each
individual topic might not 1-3 sentences descrribing
them as
opposed
to bullet points.
I like their style personally.
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
--
----------
Korey L Sewell
Graduate Student - PhD Candidate
Computer Science & Engineering
University of Michigan
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
--
----------
Korey L Sewell
Graduate Student - PhD Candidate
Computer Science & Engineering
University of Michigan
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
--
----------
Korey L Sewell
Graduate Student - PhD Candidate
Computer Science & Engineering
University of Michigan
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
--
----------
Korey L Sewell
Graduate Student - PhD Candidate
Computer Science & Engineering
University of Michigan
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
--
----------
Korey L Sewell
Graduate Student - PhD Candidate
Computer Science & Engineering
University of Michigan
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev
