On Wednesday 20 April 2011 21:55:51, Dan Doel wrote:
It's not a statistics bug. I'm reproducing it here using just vector-algorithms.
Yep. Attached a simple testcasewhich reproduces it and uses only vector and vector-algorithms.
Fill a vector of size N with [N..1], and (intro) sort it, and you get NaNs. But only with -O or above.
However, for me the NaNs disappear with the -msse2 option.
Without optimization it doesn't happen (and nothing seems to be reading/writing out of bounds, as I compiled vector with UnsafeChecks earlier and it didn't complain).
Nor does it happen here with 7.0.2 or 7.0.1.
Filling the vector with [1..N] also doesn't trigger the NaNs. [0,0..0] and [0,0..1] trigger it.
I don't know what's going on yet. I have trouble believing it's a bug in vector-algorithms code, though, as I don't think I've written any RULEs (just INLINEs), and that's the one thing that comes to mind in library code that could cause a difference between -O0 and -O. So I'd tentatively suggest it's a vector, base or compiler bug.
The above testing is on 64-bit windows running a 32-bit copy of GHC, for reference.
32-bit linux here
My ability to investigate this will be a bit limited for the near future. If someone definitively tracks it down to bugs in my code, though, let me know, and I'll try and push a new release up on hackage.
-- Dan