SIMD benchmark

[Iain Buclaw]

On 17 January 2012 08:17, Manu <turkeyman at gmail.com> wrote:
> On 17 January 2012 03:56, Iain Buclaw <ibuclaw at ubuntu.com> wrote:
>>
>> On 16 January 2012 22:36, Iain Buclaw <ibuclaw at ubuntu.com> wrote:
>> > On 16 January 2012 21:57, Peter Alexander <peter.alexander.au at gmail.com>
>> > wrote:
>> >> On 16/01/12 8:56 PM, Iain Buclaw wrote:
>> >>>
>> >>> On 16 January 2012 19:25, Walter Bright<newshound2 at digitalmars.com>
>> >>>  wrote:
>> >>>>
>> >>>> On 1/16/2012 11:16 AM, Iain Buclaw wrote:
>> >>>>
>> >>>>>>
>> >>>>>> But don't worry, I'm not planning on working on that at the moment
>> >>>>>> :-)
>> >>>>>
>> >>>>>
>> >>>>> Leave that sort of optimisation for the backend to handle please.
>> >>>>> ;-)
>> >>>>
>> >>>>
>> >>>>
>> >>>> Of course.
>> >>>>
>> >>>> I suspect Intel's compiler does that one, does gcc?
>> >>>>
>> >>>
>> >>> There's auto-vectorisation for for(), foreach(), and foreach_reverse()
>> >>> loops that I have written support for.  I am not aware of GCC
>> >>> vectorising anything else.
>> >>>
>> >>> example:
>> >>>
>> >>> int a[256], b[256], c[256];
>> >>> void foo () {
>> >>>   for (int i=0; i<256; i++)
>> >>>     a[i] = b[i] + c[i];
>> >>> }
>> >>>
>> >>
>> >> Unfortunately, if the function was this:
>> >>
>> >> void foo(int[] a, int[] b, int[] c) {
>> >>
>> >>  for (int i=0; i<256; i++)
>> >>    a[i] = b[i] + c[i];
>> >> }
>> >>
>> >> Then it can't vectorize due to aliasing.
>> >
>> > Compile with -fstrict-aliasing then?
>> >
>> > I could certainly play about with having this enabled by default, but
>> > I forsee there may be issues (maybe have it on for @safe code?)
>>
>> OK, have turned on strict aliasing by default for D2.  You should now
>> be able to vectorise loops that use locals and parameters. :-)
>
>
> What protects these ranges from being overlapping? What if they were sourced
> from pointers?
> Are just we to say in D that aliasing is not allowed, and 'you shouldn't do
> it'? People almost never alias intentionally, it's usually the
> most insidious of bugs. :/

D arrays have a .length property that keeps track of the length of the
array. When array bounds checking is turned on (default when not
compiling with -release) an assert is produced when you step outside
the bounds of the array.


Is this what you mean?

-- 
Iain Buclaw

*(p < e ? p++ : p) = (c & 0x0f) + '0';