#pragma clang loop - documentation

Aaron Ballman aaron.ballman at gmail.com
Wed Jun 11 16:05:37 PDT 2014


On Wed, Jun 11, 2014 at 7:00 PM, Tyler Nowicki <tnowicki at apple.com> wrote:
> Thanks for the reviews Aaron and Dmitri.
>
> Here is the final docs with all suggestions made.
>
> Tyler
>
>
>
>
> On Jun 11, 2014, at 7:51 AM, Aaron Ballman <aaron.ballman at gmail.com> wrote:
>
>>> +A vectorized loop performs multiple iterations of the original loop
>>> +simultaneously using vector instructions. The vector width is determined
>>> +automatically when vectorization is enabled. It can also be specified by
>>> +``vectorize_width(_value_)``, a width of 1 implies no vectorization.
>>
>> It's not entirely clear of what the value actually means.
>
> I added a line defining _value_ as a positive integer.

That's good information, but not really what I was after. How would a
user determine what value to put in there? It just says width must be
a positive integer, but that's not enough information for someone to
make a decision from. The same is basically true for interleave_count,
except there's a bit of information that the count is the number of
iterations of the original loop that are performed.  However, unless
someone already know what the operation is expected to do, they might
have a hard time understanding what "within a single iteration of the
interleaved loop" means in that sentence.

~Aaron



More information about the cfe-commits mailing list