#2374: MutableByteArray# is slower than Addr#
-------------------------------+--------------------------------------------
Reporter: dolio | Owner:
Type: bug | Status: new
Priority: normal | Milestone: 6.10 branch
Component: Compiler (NCG) | Version: 6.8.2
Severity: minor | Resolution:
Keywords: performance array | Difficulty: Unknown
Testcase: | Architecture: x86_64 (amd64)
Os: Linux |
-------------------------------+--------------------------------------------
Comment (by batterseapower):
I get results exactly opposite to those dolio reports when trying to
reproduce this. My times are:
{{{
-fvia-c:
Before SAT:
real 0m17.201s
After SAT:
real 0m10.498s
-fasm:
Before SAT:
real 0m14.942s
After SAT:
real 0m14.695s
}}}
So the NCG is unaffected by SATing the arr argument of reverse, while GCC
/is/ affected but it improves the code rather than making it worse.
I don't really understand why our results are so different. I've attached
the benchmarking script and source files I used in the hopes that someone
who understands the NCG better can work out why we aren't benefiting from
SAT while GCC is.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2374#comment:3>
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
