#1818: Code size increase vs. 6.6.1
----------------------+-----------------------------------------------------
Reporter: simonmar | Owner:
Type: bug | Status: new
Priority: normal | Milestone: 6.8.3
Component: Compiler | Version: 6.8.1
Severity: normal | Resolution:
Keywords: | Difficulty: Unknown
Testcase: | Architecture: Unknown
Os: Unknown |
----------------------+-----------------------------------------------------
Changes (by simonmar):
* milestone: 6.8.2 => 6.8.3
Comment:
Performance summary for 6.8.2 relative to 6.8.1 and 6.6.1. In all cases,
the libraries were compiled with -O2 and the nofib suite was compiled with
-O (this is just the default setting, I should probably redo the figures
using -O2).
* 6.8.2 produces binaries 7.5% smaller than 6.8.1, but still 13.5% larger
than 6.6.1
* Allocations are down a bit on average in 6.8.2. Some programs got
significantly better (mandel, x2n1), and some got worse,
particularly fulsom (+34% allocation +39% runtime), but this just puts
it back
to where it was with 6.6.1. We should look into this.
* Performance for nofib is roughly equal for 6.8.2 relative to 6.8.1
(+0.4% runtime)
* Compile times are down by 4%, but this figure tends to wobble quite a
bit.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/1818#comment:1>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
