3 Feb
2011
3 Feb
'11
9:41 a.m.
On 1 February 2011 16:23, Ulrich Weigand Ulrich.Weigand@de.ibm.com wrote:
Are they actually broken ? I'd be worried if that were the case. My understanding is that the existing ones are being used for the Neon intrinsics / builtins.
Yes, they're broken, for the reason Ira originally points out:
Right. The intrinsics are broken as well (I wrote a test to check that). The existing intrinsic testcases are compiled with -O0, which doesn't apply CSE (also there are no testcases that actually check the results of vzip, vuzp...).
I'll fix the patch according to your comments, add intrinsic testcases, and also fix neon_vtrn<mode>_internal.
Thanks, Ira