[llvm-commits] [PATCH] Add TargetMemoryModel

Hal Finkel hfinkel at anl.gov
Fri Jan 25 12:26:58 PST 2013



----- Original Message -----
> From: greened at obbligato.org
> To: "Hal Finkel" <hfinkel at anl.gov>
> Cc: dag at cray.com, llvm-commits at cs.uiuc.edu
> Sent: Friday, January 25, 2013 12:07:13 PM
> Subject: Re: [llvm-commits] [PATCH] Add TargetMemoryModel
> 
> Hal Finkel <hfinkel at anl.gov> writes:
> 
> >> What's TTI?
> >
> > Look at Analysis/TargetTransformInfo.h.
> 
> Ok, I will investigate!
> 
> >> I am certainly open to a better approach.  Would it be worth
> >> submitting the rest of what I have, reviewing it as-is and trying
> >> to
> >> commit it so that more people can participate in
> >> refining/reworking
> >> it?  I have an ongoing interest in these models so it won't be a
> >> dump
> >> and run.  :)
> >
> > I'm not sure; but it would be useful to see the rest of what you
> > have. This would provide context to what you're proposing.
> 
> Ok.  I was waiting for all of my testing runs to finish to make sure
> everything works.  I'll hopefully finish that today and post the
> whole
> sequence.

Sounds good, thanks!

> 
> >> > In addition, the prefetcher info should be cache-level specific.
> >> > Many systems have different prefetchers for different cache
> >> > levels.
> >> 
> >> The prefetcher information is purely configuration information for
> >> a
> >> software prefetcher.
> >
> > Okay, interesting; now I understand. Will you be contributing this
> > prefetcher as well?
> 
> That's a work in progress.  My best guess is that I might be able to
> convince our management to allow me to submit a *really* simple
> version
> of it.  I'm working on that.  I doubt I will be able to submit the
> whole
> thing.

Okay; I've also been working on software prefetching, so maybe we can combine efforts. Please get what you can.

> 
> The idea was to have a target-agnostic software prefetcher that could
> be
> configured on a target-by-target basis.  Even with a really dumb
> software prefetcher, it's not terribly difficult to make it smarter.
> 
> >> Maybe I should rework the names to emphasize that this is a
> >> configuration model for software prefetchers, not a model of any
> >> hardware prefetchers.
> >
> > Well, do be fair, you did name it TargetSoftwarePrefetcherInfo. It
> > might be better just to call it SoftwarePrefetcherInfo (because it
> > is
> > not a property of the target). Nevertheless, this should be part of
> > TTI so that targets can customize the behavior in the (new)
> > standard
> > way.
> 
> But it is a property of the target.  The suggested distances are
> specified for each target.  Hopefully this will be more clear once I
> submit the whole sequence.

I understand. But if you move this into TTI, where everything is a property of the target, the Target prefix will be redundant.

Thanks again,
Hal

> 
>                           -David
> 



More information about the llvm-commits mailing list