GHC-6.4.1 much slower than GHC-6.4

[Simon Marlow]

On 28 October 2005 13:12, Mirko Rahn wrote:

>> Can you try with -fno-full-laziness (with both compilers).  I wonder
>> if it's a "does this get floated out" issue.
> 
> Bit slower than without this flag:
> 
> 6.4:   cpu-time interesting (3,3): 00:01:40.46
> 6.4.1: cpu-time interesting (3,3): 00:06:11.13
> 
> But the ratio keeps the same.

If you can find the root of the computation that is being evaluated
multiple times with 6.4.1, that would help a great deal.  The call-tree
on the time profile should help a lot, and you can narrow it down by
inserting {-# SCC #-} annotations.  Basically we want to know which
expression is being evaluated multiple times.

Cheers,
	Simon