[Numpy-discussion] Seg. fault when running tests
Sat Jun 15 17:15:24 CDT 2013
On 15.06.2013 22:26, Charles R Harris wrote:
> On Sat, Jun 15, 2013 at 2:23 PM, Warren Weckesser
> <email@example.com <mailto:firstname.lastname@example.org>> wrote:
> Also on 2.7. The -O3 flag seems to cause the problem.
I was wrong when I changed this flag.
The compiler is allowed to assume the data is aligned to the size of the
So when vectorizing it peels the wrong amount of data and ends up with
non 16-byte aligned SSE instructions.
I'll file a pull reverting the change.
More information about the NumPy-Discussion