[GHC] #8971: Native Code Generator for 8.0.1 is not as optimized as 7.6.3...

[GHC]

#8971: Native Code Generator for 8.0.1 is not as optimized as 7.6.3...
-------------------------------------+-------------------------------------
        Reporter:  GordonBGood       |                Owner:
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler (NCG)    |              Version:  8.0.1
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Runtime           |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by GordonBGood):

 Replying to [comment:30 bgamari]:
 > I'm de-milestoning this due to a lack of developer interest. Currently
 we have many bugs and few developers; many of these bugs affect platforms
 far more common than i386 and these are the bugs we will focus on.

 I agree that this issue isn't all that important in the Great Scheme Of
 Things given that the Native Code Generator (NCG) is gradually being
 relegated to just a builder for quick-cycle debugging purposes as the
 (slower compiling, but producing generally faster code) LLVM back end
 becomes more and more stable (and usable, especially if the required LLVM
 files, including patches, are distributed with/made available for each GHC
 version starting with 8.2.1, as
 promised[https://ghc.haskell.org/trac/ghc/wiki/ImprovedLLVMBackend]).

 Much more important are performance bugs that affect '''both''' the NCG
 and LLVM back ends such as the (I'm beginning to think related to each
 other) #12798 and #12808 pair of issues that although only slowing
 performance by up to 40% or so (although for all platforms instead of only
 x86 by 100% as here), affect the perception that GHC code is "slower than
 Cee" and for which there is no workaround such as choosing an alternate
 back end.

 I'll monitor this and will probably either close the issue or at least
 reduce its priority if version 8.2.1 integrates the LLVM back end as well
 as it promises.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8971#comment:31>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler