<div dir="ltr"><div>What about types like i24 that likely need 2 or more loads with shifts? I imagine synthesizing an i24 constant in a register might be simpler than that?</div><div><br></div><div>Are you suggesting to remove the check altogether?</div><br clear="all"><div><div dir="ltr" data-smartmail="gmail_signature">~Craig</div></div><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Jul 30, 2021 at 11:59 AM Eli Friedman <<a href="mailto:efriedma@quicinc.com" target="_blank">efriedma@quicinc.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div lang="EN-US">
<div>
<p class="MsoNormal">This logic was originally added in <a href="https://reviews.llvm.org/rG65df808f6254617b9eee931d00e95d900610b660" target="_blank">
https://reviews.llvm.org/rG65df808f6254617b9eee931d00e95d900610b660</a> . I’m not sure what the code is trying to do.<u></u><u></u></p>
<p class="MsoNormal"><u></u> <u></u></p>
<p class="MsoNormal">In general, the number of instructions it takes to load a value into registers from memory should be comparable to the number of instructions it takes to synthesize it in registers. It shouldn’t really matter if the type is legal. I can
think of two reasons the type would matter:<u></u><u></u></p>
<p class="MsoNormal"><u></u> <u></u></p>
<ol style="margin-top:0in" start="1" type="1">
<li style="margin-left:0in">Controlling the size of the lookup table; creating a table with types that are very large, or have a lot of padding, could be inefficient.<u></u><u></u></li><li style="margin-left:0in">Lowering a PHI node here could make it harder for other passes, like instcombine, to slice it up later. But we’re restricted the switch-to-lookup table combine so it only runs late,
so probably less of a concern now than it used to be.<u></u><u></u></li></ol>
<p class="MsoNormal"><u></u> <u></u></p>
<p class="MsoNormal">-Eli<u></u><u></u></p>
<p class="MsoNormal"><u></u> <u></u></p>
<div style="border-top:none;border-right:none;border-bottom:none;border-left:1.5pt solid blue;padding:0in 0in 0in 4pt">
<div>
<div style="border-right:none;border-bottom:none;border-left:none;border-top:1pt solid rgb(225,225,225);padding:3pt 0in 0in">
<p class="MsoNormal"><b>From:</b> Craig Topper <<a href="mailto:craig.topper@gmail.com" target="_blank">craig.topper@gmail.com</a>> <br>
<b>Sent:</b> Friday, July 30, 2021 8:56 AM<br>
<b>To:</b> llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>>; Eli Friedman <<a href="mailto:efriedma@quicinc.com" target="_blank">efriedma@quicinc.com</a>>; Alex Bradbury <<a href="mailto:asb@lowrisc.org" target="_blank">asb@lowrisc.org</a>><br>
<b>Subject:</b> [EXT] SimplifyCFG's switch to lookup table with illegal types.<u></u><u></u></p>
</div>
</div>
<p class="MsoNormal"><u></u> <u></u></p>
<div>
<div>
<p class="MsoNormal">ShouldBuildLookupTable contains a check to make sure the type of the Phi is legal for the target. This currently prevents i32 lookup tables from being formed on RISCV64 which only has i64 as a legal type. Obviously i32 is going to be a
more widespread type than i64 for C code so I would like to improve this. But i8/i16 are also disabled on RISCV, ARM, AArch64, and probably other targets which can do an i8/i16 load.<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal"><u></u> <u></u></p>
</div>
<div>
<p class="MsoNormal">What is the right check here? Can we make sure it fits in the largest native integer type from data layout instead? Or should we add another TTI hook to ask the target?<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal"><u></u> <u></u></p>
</div>
<p class="MsoNormal">Thanks,<br clear="all">
<u></u><u></u></p>
<div>
<div>
<p class="MsoNormal">~Craig<u></u><u></u></p>
</div>
</div>
</div>
</div>
</div>
</div>
</blockquote></div>