[llvm-dev] SelectionDAG::LegalizeTypes is very slow in 3.1 version

xiaoyong liu via llvm-dev llvm-dev at lists.llvm.org
Tue Sep 27 10:07:54 PDT 2016


Absolutely, James. I was with LLVM from very beginning. And keep everything up-to-date.

The background is the new product was inherited. It took time to upgrade.


To the problem: If anyone meet the similar issue, pls kindly share.


________________________________
From: James Y Knight <jyknight at google.com>
Sent: Tuesday, September 27, 2016 8:25:55 AM
To: xiaoyong liu
Cc: llvm-dev at lists.llvm.org
Subject: Re: [llvm-dev] SelectionDAG::LegalizeTypes is very slow in 3.1 version

You realize 3.1 was released **4 years ago**, and that there's been more than 50,000 commits to llvm since then (not even including commits to clang and other pieces!), right?


On Tue, Sep 27, 2016 at 2:20 AM, xiaoyong liu via llvm-dev <llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>> wrote:

In 3.1, the backend is very slow to legalize types.


Following is the code snippet which may be the culprit:

  %Result.i.i.i97 = alloca i33, align 8
  %Result.i.i.i96= alloca i33, align 8
  %Result.i.i.i95 = alloca i33, align 8
  %Result.i.i.i94 = alloca i33, align 8
  %Result.i.i.i93 = alloca i33, align 8
  %Result.i.i.i92= alloca i33, align 8
  %Result.i.i.i91 = alloca i33, align 8
  %Result.i.i.i90 = alloca i33, align 8
  %Result.i.i.i89 = alloca i33, align 8


The compilation time improve significant if running on LLVM 3.9. Unfortunately, it's not a choice for me now to upgrade it to 3.9.

If anyone luckily to  know what's the evolution in this area to improve the compilation time performance, pls kindly comments.


thanks

Xiaoyong


_______________________________________________
LLVM Developers mailing list
llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>
http://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/20160927/4afb12ea/attachment.html>


More information about the llvm-dev mailing list