@carter: Yes; I think @rwbarton is quite right in his analysis!
@rwbarton: Thanks for nailing this one!
I'd consider this a bug; but I'm not sure how important it is to reserve
cycles. @simonpj can definitely make that call. Maybe it can be merged with
@carter's ticket #9276, and addressed within that context.
On Mon, Jul 14, 2014 at 10:17 AM, GHC
#9304: Floating point woes; Different behavior on 32- vs 64-bit x86 -------------------------------------+------------------------------------ Reporter: lerkok | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.3 Resolution: | Keywords: floating point Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: 9276 Blocking: | Related Tickets: -------------------------------------+------------------------------------
Comment (by carter):
@lerkok I think @rwbarton has the problem nailed.
If I'm understanding everything correctly, the culprit here is that GHC's const evaluation tooling doesn't respect the implicit floating point model induced by -msse2 vs -f-excess-precision ?
-- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9304#comment:29 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler