[GHC] #9315: Weird change in allocation numbers of T9203

[GHC]

#9315: Weird change in allocation numbers of T9203
-------------------------------------+-------------------------------------
              Reporter:  nomeata     |             Owner:  simonmar
                  Type:  bug         |            Status:  new
              Priority:  normal      |         Milestone:
             Component:  Runtime     |           Version:  7.9
  System                             |          Keywords:
            Resolution:              |  Operating System:  Unknown/Multiple
Differential Revisions:              |   Type of failure:  None/Unknown
          Architecture:              |         Test Case:
  Unknown/Multiple                   |          Blocking:
            Difficulty:  Unknown     |
            Blocked By:              |
       Related Tickets:              |
-------------------------------------+-------------------------------------

Comment (by nomeata):

 > validate takes a minute longer

 Out of more than half an hour. It’s not like validate is fast enough so
 that people can actually sit and watch it. And no matter whether its 36 or
 37 minutes, I guess most will stop running validate themselves and rather
 leave that to Phabricator or other CI stuff, precisely because validate
 takes so long and its convenient to let someone else do it for you.

 > the tests now fail with a quick build (which we recommend for
 development, and I generally use)

 True. But development trees are inherently unusable, as people might have
 all kinds of settings there (`-O`, `-DDEBUG` etc.) that invalidate
 performance numbers. I never paid attention to them for that reason,
 unless I was doing a clean validate run in a separate checkout.

 Isn’t running validate in the development tree discouraged anyways?

 > the compiler is still not built with -O2

 True, but one step at a time. (I’ll time this change later, probably over
 lunch.)


 You say `quick` is recommended; I thought `devel2` is recommended and
 that’s what I use and recommend. And there will be much more variation in
 people’s development trees settings, it’s hard to aim for them. What we
 can aim for are our two well-defined settings: The default (i.e. what we
 release) and validate. I’m arguing that we should unify these two as much
 as possible.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9315#comment:21>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler