[llvm-dev] [RFC] Removing Waymarking from Use.

Sat Apr 18 07:06:32 PDT 2020

>From looking at UseDefLists.h, it seems generic enough to reimplement Uses in llvm with it.

but I don’t think we are gaining much by doing it. Why do you think it is “clearly it is superior to `llvm::Use`” ?

________________________________
From: Ehud Katz <ehudkatz at gmail.com>
Sent: Friday, April 17, 2020 10:25 PM
To: Eric Christopher <echristo at gmail.com>
Cc: Tyker1 at outlook.com <Tyker1 at outlook.com>; llvm-dev at lists.llvm.org <llvm-dev at lists.llvm.org>
Subject: Re: [llvm-dev] [RFC] Removing Waymarking from Use.

Now that D77144 has landed; any thoughts regarding what I suggested? Using UseDefLists.h?

On Wed, 15 Apr 2020 at 15:32 Ehud Katz <ehudkatz at gmail.com<mailto:ehudkatz at gmail.com>> wrote:
Maybe we can utilize the implementation in mlir/IR/UseDefLists.h<https://github.com/llvm/llvm-project/blob/master/mlir/include/mlir/IR/UseDefLists.h> in here (clearly it is superior to `llvm::Use`) ?
By that we will have the same code base (instead of duplicate implementations of Use-Lists).
D77144<https://reviews.llvm.org/D77144> should definitely go in, first, though.

On Tue, Apr 14, 2020 at 8:32 PM Eric Christopher via llvm-dev <llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>> wrote:
Yes please.

On Tue, Apr 14, 2020, 5:02 AM Tyker1 at outlook.com<mailto:Tyker1 at outlook.com> via llvm-dev <llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>> wrote:
a bit of time has passed and there to my knowledge still no strong arguments against removing it.
is everyone OK to proceed with the removal ?

Gauthier
________________________________
From: Chris Lattner <clattner at nondot.org<mailto:clattner at nondot.org>>
Sent: Saturday, April 4, 2020 7:40 PM
To: Johannes Doerfert <johannesdoerfert at gmail.com<mailto:johannesdoerfert at gmail.com>>
Cc: Ehud Katz <ehudkatz at gmail.com<mailto:ehudkatz at gmail.com>>; llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org> <llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>>; Tyker1 at outlook.com<mailto:Tyker1 at outlook.com> <Tyker1 at outlook.com<mailto:Tyker1 at outlook.com>>
Subject: Re: [llvm-dev] [RFC] Removing Waymarking from Use.

On Apr 3, 2020, at 11:06 AM, Johannes Doerfert <johannesdoerfert at gmail.com<mailto:johannesdoerfert at gmail.com>> wrote:

Is it worth it? I think it is. But I am not sure I see the whole picture -
are there low-memory systems that need to run LLVM on?

I am not sure what needs to be done to approve such a fundamental change;
especially when we can't prove the Waymarking was needed at all.

I guess if no-one brings forth arguments (= results) for keeping it and

people continue to support replacing it, we will replace it. There should

be a grace period in which people have the chance to do their benchmarking

(basically what is happening), but I don't recall a problem being reported yet.

I agree.  I’m not hearing strong arguments to retain it, so let's remove it.  Worst case, we can always reinstate it if there is a good reason discovered down the line.  Thank you!

-Chris
_______________________________________________
LLVM Developers mailing list
llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
_______________________________________________
LLVM Developers mailing list
llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20200418/9b030bca/attachment.html>