<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">Hi Lang,<div class=""><br class=""></div><div class="">Thanks for the clarifications.</div><div class=""><br class=""><div><blockquote type="cite" class=""><div class="">On Mar 9, 2015, at 2:30 PM, Lang Hames <<a href="mailto:lhames@gmail.com" class="">lhames@gmail.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div dir="ltr" class="">Hi Quentin, Jonas,<div class=""><br class=""></div><div class="">Splitting fits in with PBQP reasonably well, at least conceptually. The PBQP graph is designed to be mutable, so there is no problem with updating it when splitting.</div><div class=""><br class=""></div><div class="">As I see it, there are two logical places to integrate splitting into PBQP:</div><div class=""><br class=""></div><div class="">1) Split during spilling -- If a PBQP solution selects the spill option for a node, rather than spill immediately, split the interval first. This seems (superficially) like an opportunity to use trySplit, but I haven't looked at the code for it. If trySplit is looking at the existing allocation state to make those decisions we may have trouble: In RAGreedy some "current" colorings may be revisited, but during spilling in PBQP we're guaranteed that *all* colorings will be revisited in the next round, which may throw trySplit's heuristics out.</div></div></div></blockquote><div><br class=""></div><div>trySplit splits either the given virtual register or some of its interference to make it colorable based indeed on the current state of allocation. That being said, I do not know if would really be a problem if you would revisited everything in the next round. I guess it would depend how you handle coalescing for those new split points.</div><div><br class=""></div><div>I thought the PBQP was not revisiting everything that is why I believed trySplit as it is what not a good candidate for splitting in PBQP since it requires you reconsider some of the nodes you already colored. But now, I remember the IRC approach and yeah I believe it would work (modulo the coalescing of the new split points and the cost of rebuild/update part of the graph).</div><br class=""><blockquote type="cite" class=""><div class=""><div dir="ltr" class=""><div class=""><br class=""></div><div class="">2) Pre-split -- We could have a pre-pass split live ranges at sensible points (e.g. around loops) before we enter the coloring/spilling loop. PBQP tries to coalesce as it colors, so it will try to undo these splits where possible. I assume trySplit would not be a good fit here, but we may be able to reuse other parts of the splitting code.</div></div></div></blockquote><div><br class=""></div><div>I agree.</div><div><br class=""></div><div>Cheers,</div><div>-Quentin</div><br class=""><blockquote type="cite" class=""><div class=""><div dir="ltr" class=""><div class=""><br class=""></div><div class="">We do want to add splitting support to PBQP, so we should be on the lookout for opportunities to share code where it makes sense.</div><div class=""><br class=""></div><div class="">Cheers,</div><div class="">Lang.</div><div class=""> </div></div><div class="gmail_extra"><br class=""><div class="gmail_quote">On Sat, Mar 7, 2015 at 5:01 AM, Quentin Colombet <span dir="ltr" class=""><<a href="mailto:qcolombet@apple.com" target="_blank" class="">qcolombet@apple.com</a>></span> wrote:<br class=""><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word" class="">Hi Jonas,<div class=""><br class=""><div class=""><span class=""><blockquote type="cite" class=""><div class="">On Mar 6, 2015, at 12:31 AM, Jonas Paulsson <<a href="mailto:jonas.paulsson@ericsson.com" target="_blank" class="">jonas.paulsson@ericsson.com</a>> wrote:</div><br class=""><div class=""><div style="font-family:Helvetica;font-size:12px;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px" class=""><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class="">Hi,<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class="">I have worked a little on the PBQP register allocator, and it is quite clear (at least to me) that it is not even a serious alternative to RegAllocGreedy at the moment, due to the poor handling of spilling. As Arnaud wrote below, it is not optimizing spilling at all, but rather just spills anything that does not get an assignment. The result is a lot more spill/reload instructions than needed.<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class="">In RegAllocBase.h it says “…Register allocation complexity, and generated code performance is determined by the effectiveness of live range splitting rather than optimal coloring…”. I would then think that any register allocation algorithm should benefit from this, but find that only RegAllocGreedy is doing live range splitting, and that the code for doing this is local to that allocator.<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class="">I would like to suggest a refactoring to make RAGreedy::trySplit() and its sub functions callable from any register allocator. Perhaps part of SplitEditor?</span></div></div></div></blockquote><div class=""><br class=""></div></span><div class="">What do you expect from sure refactoring?</div><div class=""><br class=""></div><div class="">In the current form, live-range splitting for the PBQP implies to rebuild part of the graph and I suspect it would be easier to rebuild it from scratch than trying to update it if we would want to use it.</div><div class="">I believe that to have an efficient implementation, the PBQP splitting should work directly on the graph and not on the program as it is the case for the Greedy Allocator.</div><div class=""><br class=""></div><div class="">Now, regarding the fast register allocator, I am not sure it could cope at all with live-range splitting as it would change some allocation decisions that were supposed to be final.</div><div class=""><br class=""></div><div class="">The bottom line is I do not believe there is much to share here.</div><div class=""><br class=""></div><div class="">Cheers,</div><div class="">-Quentin</div><span class=""><br class=""><blockquote type="cite" class=""><div class=""><div style="font-family:Helvetica;font-size:12px;font-style:normal;font-variant:normal;font-weight:normal;letter-spacing:normal;line-height:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px" class=""><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class=""><u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class="">What do you think about this?<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class="">/Jonas<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span style="color:rgb(31,73,125)" class=""> </span></div><div class=""><div style="border-style:solid none none;border-top-color:rgb(181,196,223);border-top-width:1pt;padding:3pt 0cm 0cm" class=""><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><b class=""><span style="font-size:10pt;font-family:Tahoma,sans-serif" class="">From:</span></b><span style="font-size:10pt;font-family:Tahoma,sans-serif" class=""><span class=""> </span>Arnaud A. de Grandmaison [<a href="mailto:arnaud.degrandmaison@arm.com" target="_blank" class="">mailto:arnaud.degrandmaison@arm.com</a>]<span class=""> </span><br class=""><b class="">Sent:</b><span class=""> </span>den 4 mars 2015 15:43<br class=""><b class="">To:</b><span class=""> </span>Jonas Paulsson; Lang Hames<br class=""><b class="">Cc:</b><span class=""> </span><a href="mailto:llvmdev@cs.uiuc.edu" target="_blank" class="">llvmdev@cs.uiuc.edu</a><br class=""><b class="">Subject:</b><span class=""> </span>RE: PBQP spilling<u class=""></u><u class=""></u></span></div></div></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><u class=""></u> <u class=""></u></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span lang="EN-GB" style="color:rgb(31,73,125)" class="">Yes, for now the spilling is done in the most basic way, i.e. it’s functionally correct --- but not efficient. The focus was on the allocator itself, not on the spilling. As you noticed, the work still to be done in this area is live range splitting, and smarter spill code insertion. Another area is improving the reduction order, to make the allocator less sensitive to the reduction order.<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span lang="EN-GB" style="color:rgb(31,73,125)" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span lang="EN-GB" style="color:rgb(31,73,125)" class="">There is no official plan; we started to discuss that with Lang some time ago, but none of us had time to dive into it yet. Any help appreciated there<span class=""> </span></span><span lang="EN-GB" style="font-family:Wingdings;color:rgb(31,73,125)" class="">J</span><span lang="EN-GB" style="color:rgb(31,73,125)" class="">.<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span lang="EN-GB" style="color:rgb(31,73,125)" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span lang="EN-GB" style="color:rgb(31,73,125)" class="">Cheers,<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span lang="EN-GB" style="color:rgb(31,73,125)" class="">Arnaud<u class=""></u><u class=""></u></span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span lang="EN-GB" style="color:rgb(31,73,125)" class=""> </span></div><div style="border-style:none none none solid;border-left-color:blue;border-left-width:1.5pt;padding:0cm 0cm 0cm 4pt" class=""><div class=""><div style="border-style:solid none none;border-top-color:rgb(181,196,223);border-top-width:1pt;padding:3pt 0cm 0cm" class=""><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><b class=""><span style="font-size:10pt;font-family:Tahoma,sans-serif" class="">From:</span></b><span style="font-size:10pt;font-family:Tahoma,sans-serif" class=""><span class=""> </span>Jonas Paulsson [<a href="mailto:jonas.paulsson@ericsson.com" style="color:purple;text-decoration:underline" target="_blank" class="">mailto:jonas.paulsson@ericsson.com</a>]<span class=""> </span><br class=""><b class="">Sent:</b><span class=""> </span>04 March 2015 13:51<br class=""><b class="">To:</b><span class=""> </span>Lang Hames; Arnaud De Grandmaison<br class=""><b class="">Cc:</b><span class=""> </span><a href="mailto:llvmdev@cs.uiuc.edu" style="color:purple;text-decoration:underline" target="_blank" class="">llvmdev@cs.uiuc.edu</a><br class=""><b class="">Subject:</b><span class=""> </span>PBQP spilling<u class=""></u><u class=""></u></span></div></div></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><span lang="EN-GB" class=""> </span></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class="">Hi,<u class=""></u><u class=""></u></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><u class=""></u> <u class=""></u></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class="">I would like to ask about PBQPs use of InlineSpiller. The code output when using PBQP gets a lot bigger compared to when using RegAllocGreedy. PBQP does not split the live intervals, and a lot more (often redundant) reload instructions are emitted as a result, it seems. I wonder why this is, and if there are any plans to improve on this point?<u class=""></u><u class=""></u></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class=""><u class=""></u> <u class=""></u></div><div style="margin:0cm 0cm 0.0001pt;font-size:11pt;font-family:Calibri,sans-serif" class="">/Jonas Paulsson</div></div></div></div></blockquote></span></div><br class=""></div></div></blockquote></div><br class=""></div>
</div></blockquote></div><br class=""></div></body></html>