<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<p><br>
</p>
<div class="moz-cite-prefix">On 04/25/2017 11:32 PM, Daniel Berlin
via llvm-dev wrote:<br>
</div>
<blockquote
cite="mid:CAF4BwTVWWNnUTH66QdEDFFD1m=UPhWmQeEFYTKKBpKAfXprohQ@mail.gmail.com"
type="cite">
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<div dir="ltr">For those not following along, startingin r301379,
Simplify* in InstructionSimplify now can just take a query
struct instead of 8000 optional arguments. Nothing is really
new since it used the same thing under the covers.
<div>
<div><br>
</div>
<div>I'm slowly converting the old uses away (deletion of the
old APIs is a different question).</div>
<div>Staring at most of them, i could just directly convert
them using braced list initialization, and construct the
objects again and again, but the vast majority of uses
actually:</div>
<div><br>
</div>
<div>A. don't change the relevant query arguments over the
lifetime of the pass</div>
<div>B. actually have more information sitting around than
they are passing along.</div>
<div><br>
</div>
<div>For example, instcombine has a DomTree and
AssumptionCache that are required, but it doesn't pass them
to SimplifyBinOp in a bunch of cases.</div>
<div><br>
</div>
<div>JumpThreading has TLI but doesn't pass it to
SimplifyCmpInst.</div>
<div><br>
</div>
<div>CVP at least admits it has a problem:</div>
<div>"CorrelatedValuePropagation.cpp: // FIXME: Provide TLI,
DT, AT to SimplifyInstruction.</div>
<div>CorrelatedValuePropagation.cpp: if (Value *V =
SimplifyInstruction(P, DL)) {</div>
<div>"</div>
<div>(This is because it uses LVI, which requires those
things, but it doesn't ask for them itself)</div>
<div><br>
</div>
<div>Assuming this is not deliberate, it would seem to me to
just be easiest to set up the query structure once in the
pass and use it everywhere, which would hopefully avoid
these issues in the future, besides making most of the call
strings dramatically shorter :)</div>
<div><br>
</div>
<div>Is there any real downside (compile time performance,
etc) to this vs what we do now (which is basically
constructing the query objects again and again under the
covers) that anyone can think of?</div>
<div><br>
</div>
<div>(again, assuming we are not deliberately avoiding passing
info to Simplify* in most of these cases)</div>
</div>
</div>
</blockquote>
<br>
I don't think it is deliberate; it is just that no one has done the
work to make it uniform. It is possible that doing this will expose
some compile-time issues, but if so, we should deal with them for
real (not just hide them by a failure to provide
optional-but-available analyses).<br>
<br>
I'm certainly in favor of more uniformity here.<br>
<br>
-Hal<br>
<br>
<blockquote
cite="mid:CAF4BwTVWWNnUTH66QdEDFFD1m=UPhWmQeEFYTKKBpKAfXprohQ@mail.gmail.com"
type="cite">
<div dir="ltr">
<div>
<div><br>
</div>
</div>
<div>I figured i'd ask before i went writing the scripts/etc to
help convert a bunch of this stuff :)</div>
<div><br>
</div>
</div>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
LLVM Developers mailing list
<a class="moz-txt-link-abbreviated" href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>
<a class="moz-txt-link-freetext" href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a>
</pre>
</blockquote>
<br>
<pre class="moz-signature" cols="72">--
Hal Finkel
Lead, Compiler Technology and Programming Languages
Leadership Computing Facility
Argonne National Laboratory</pre>
</body>
</html>