<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body>
<div class="moz-text-flowed" style="font-family: -moz-fixed;
font-size: 12px;" lang="x-unicode">Hi all,
<br>
<br>
I am looking into the benefits of a VPlan-based cost model, and
into how such a cost model should be implemented to make the most
out of these benefits.
<br>
<br>
Over the last year I have been working with LLVM, mainly focused
on the ARM backend, in the course of a one-year internship at Arm
Ltd. My main project from December 2019 to August 2020 was to
introduce gather/scatters for MVE auto-vectorization. One of the
recurring challenges during this work was to get things right in
the cost model.
<br>
For example, gathers can extend the data they load, while scatters
can truncate their input, meaning that an extend following the
load, or a truncate preceding the store, is for free if it meets
certain type conditions. As the current cost model is not designed
for context-based analysis, this was a pain to model.
<br>
<br>
I have done some research and found out that one of the proposed
benefits of VPlan is that a new cost model making use of it would
be able to better support context-dependent decisions like this.
<br>
However, there does not exist much specification about what such a
cost model should look like.
<br>
<br>
Also, I have read through the respective code to understand how
loop vectorization is currently done and how far the introduction
of VPlan has progressed and have realised that the only recipe
that actually handles more than one instruction from the input IR
is the one for interleaved groups. When the VPlan is generated on
the VPlan-native path, every IR instruction is considered and
transformed into a recipe separately, ignoring its context (to
give a code location, I am looking at
VPlanTransforms::VPInstructionsToVPRecipes).
<br>
And maybe there are architectures that for some cases do not have
the same vector instructions, so a pattern that works great for
one could be useless for others. So I am wondering: Is there any
plan to have target-dependent flavours of recipes, or how will
those things be handled?
<br>
Right now it makes sense that nothing like this has been
implemented yet, as there is no cost model that could guide the
transformation. But if recipes are a general thing, should the
cost model be the component actually providing the target-specific
pattern for a recipe, together with its cost?
<br>
<br>
I am considering choosing a topic related to VPlan, possibly cost
modelling, for my Master thesis, with the goal to present a
solution and implement a prototype.
<br>
<br>
What I would like to ask the community is
<br>
<br>
(1) what goals there are or should be for VPlan-based cost
modelling,
<br>
(2) whether there have been any discussions on target-dependent
patterns yet, and
<br>
(3) which examples of inefficiencies and shortcomings of the
current cost model they have come across.
<br>
<br>
I am looking forward to your feedback.
<br>
<br>
Many thanks,
<br>
Anna Welker
<br>
<br>
</div>
</body>
</html>