[GHC] #6104: Regression: space leak in HEAD vs. 7.4
GHC
cvs-ghc at haskell.org
Mon Jul 16 17:35:33 CEST 2012
#6104: Regression: space leak in HEAD vs. 7.4
---------------------------------+------------------------------------------
Reporter: simonmar | Owner: simonpj
Type: bug | Status: new
Priority: highest | Milestone: 7.6.1
Component: Compiler | Version: 7.5
Keywords: | Os: Unknown/Multiple
Architecture: Unknown/Multiple | Failure: Compile-time performance bug
Difficulty: Unknown | Testcase:
Blockedby: | Blocking:
Related: |
---------------------------------+------------------------------------------
Comment(by simonpj):
Interface file size:
{{{
GHC 7.4.1 -O -rw-rw-r-- 1 simonpj GHC 5983 Jul 16 15:21 T6104.hi
HEAD -rw-rw-r-- 1 simonpj GHC 10487 Jul 16 14:08 T6104.hi
}}}
So, yes, the interface file is a lot bigger. Looking more closely, this
seems largely accidental. The code in HEAD is broken into more functions,
each of which is individually small enough to be below the inlining-
creation threshold.
I compiled Cabal with -O, and compared the total size (with `ls -l`) of
the `.o` files and `.hi` files with both compilers:
{{{
HEAD change over 7.4.1
.o file size -6%
.hi file size +8%
}}}
So yes, interface file size is up, but not massively, and .o file size is
down.
This does not seem to account for the originally-reported doubling of
residency! And (I've checked) that doubling is still happening.
Conclusion: I don't think it's just bigger interface files. I there is
some other space leak too.
Simon
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/6104#comment:9>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
More information about the Glasgow-haskell-bugs
mailing list