Hi Richard,
On 8 November 2017 at 20:11, Jeff Law wrote:
> On 11/03/2017 10:18 AM, Richard Sandiford wrote:
>> This patch adds a routine that lists the available vector sizes
>> for a target and uses it for some existing target conditions.
>> Later patches add more uses.
>>
>> The cases are taken from multiple_sizes.
>>
>>
>> 2017-11-03 Richard Sandiford
>> Alan Hayward
>> David Sherwood
>>
>> gcc/testsuite/
>> * lib/target-supports.exp (available_vector_sizes): New proc.
>> (check_effective_target_vect_multiple_sizes): Use it.
>> (check_effective_target_vect64): Likewise.
>> (check_effective_target_vect_sizes_32B_16B): Likewise.
This patch (r254590) means that we now skip:
gcc.dg/vect/vect-33.c -flto -ffat-lto-objects
scan-tree-dump-times vect "Alignment of access forced using
versioning" 1
gcc.dg/vect/vect-33.c scan-tree-dump-times vect "Alignment of
access forced using versioning" 1
on armeb-none-linux-gnueabihf --with-fpu=neon-fp16
This is because of the change in vect64, which used to give a
different result on arm depending on the endianness.
I think that's not a big deal, but I thought I should mention it.
Christophe
> OK.
> jeff