Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.c++ > #32875

Re: Why compiler is able to vectorize on int but not on unsigned

Cross-posted to 2 groups.

Show key headers only | View raw

On Thu, 04 Dec 2014 22:04:33 +0100
Alain Ketterlin <alain@dpt-info.u-strasbg.fr> wrote:

> James Kuyper <jameskuyper@verizon.net> writes:
> 
> > On 12/03/2014 12:53 PM, Alain Ketterlin wrote:
> >> Melzzzzz <mel@zzzzz.com> writes:
> >> 
> >>> int testNormal(const vector_t & ghs, const vector_t & lhs) {
> >>>     int max = 0;
> >>>     int tempMax;
> >>>     for (unsigned k = 0; k < ghs.size(); k++) {
> >>>         tempMax = lhs[k] + ghs[k];
> >>>         if (max < tempMax) {
> >>>             max = tempMax;
> >>>         }
> >>>     }
> >>>     return max;
> >>> }
> >>>
> >>> where: typedef vector<int> vector_t;
> >>> (could be also plain array so I post this to clc as well)
> >> 
> >> A plain array would have been better, because the explanation has
> >> nothing to do with vectors/templates/whatnot
> 
> Sorry for the late answer, I was busy.
> 
> > He's already posted (nearly three hours ago) a message indicating
> > that the problem does not occur when using plain arrays,
> 
> You're right. Some people stop reading in the middle of posts when
> they meet a template. I admit I did stop reading in the middle of the
> thread for lack of time.
> 
> (I come back to plain arrays below.)
> 
> > so the explanation must indeed have something to do with
> > vectors/templates/whatnot - to be more precise, it seems to be
> > connected to the use of ghs.size().
> 
> And your suggestion is?
> 
> Here is mine: ghs.size() (probably) has return type size_t, which is
> (probably) bigger than unsigned on the target machine. (Actually
> size() return size_type which is typedefed inside vector, usually as
> size_t bt I'm not sure this is mandatory.)
> 
> Therefore, the loop may never end (if k wraps around---which it is
> allowed to when unsigned), and this is enough for the compiler to
> decide *not* to vectorize (because manipulating the loop trip count
> in that case is tricky). The same thing does not happen when k is
> signed: this is explained in the link I posted earlier.
> 
> So, why does it work with plain arrays? Well, it does not, but
> Melzzzzz's modified program was vectorized because the size was now an
> unsigned (not a size_t) coming in as a parameter, and the risk of
> overflow during iteration was gone.
> 
> So why do I say it does not work with a "plain-array-like" version of
> vector? Because a (rough) translation of vector<int> using a plain
> array would be a structure like:
> 
> struct vectorlike {
>     size_t sz;
>     int * data;
> };
> 
> and bam! the same would happen in C if the sz field were used as a
> loop bound. Also, in the version of Melzzzzz if the parameters were
> "real" plain arrays but the size were passed in as a size_t (instead
> of unsigned, as it should I think), the same would happen. (And this
> make the cross-posting to comp.lang.c not completely irrelevant.)

You are absolutely right. If I change function
to:
int testNormal(const int* ghs, const int* lhs,size_t sz) {
    int max = 0; 
    int tempMax; 
    for (unsigned k = 0; k < sz; k++) {
        tempMax = lhs[k] + ghs[k];
        if (max < tempMax) {
            max = tempMax;  
        }
    }
    return max;
}

it *does not* vectorize as well!

> 
> I'm speculating, because I haven't looked into the respective
> vectorizers, but my experience with that unsigned/signed mess in
> compilers makes me reasonably sure the problem lies somewhere there. I
> maintain what I said, but of course if James or anybody else has a
> better explanation I would love to read it.

Your explanation fits what happens with the code. It has to do
with unsigned is smaller then size_t and can overflow.


> 
> -- Alain.

Back to comp.lang.c++ | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Why compiler is able to vectorize on int but not on unsigned Melzzzzz <mel@zzzzz.com> - 2014-12-03 11:19 +0100
  Re: Why compiler is able to vectorize on int but not on unsigned James Kuyper <jameskuyper@verizon.net> - 2014-12-03 08:02 -0500
    Re: Why compiler is able to vectorize on int but not on unsigned jacob navia <jacob@spamsink.net> - 2014-12-03 14:57 +0100
      Re: Why compiler is able to vectorize on int but not on unsigned Tim Prince <tprince@computer.org> - 2014-12-03 09:23 -0500
        Re: Why compiler is able to vectorize on int but not on unsigned James Kuyper <jameskuyper@verizon.net> - 2014-12-03 10:03 -0500
      Re: Why compiler is able to vectorize on int but not on unsigned Victor Bazarov <v.bazarov@comcast.invalid> - 2014-12-03 09:24 -0500
        Re: Why compiler is able to vectorize on int but not on unsigned Richard Heathfield <invalid@see.sig.invalid> - 2014-12-03 14:36 +0000
    Re: Why compiler is able to vectorize on int but not on unsigned Melzzzzz <mel@zzzzz.com> - 2014-12-03 16:03 +0100
      Re: Why compiler is able to vectorize on int but not on unsigned Victor Bazarov <v.bazarov@comcast.invalid> - 2014-12-03 10:27 -0500
        Re: Why compiler is able to vectorize on int but not on unsigned Melzzzzz <mel@zzzzz.com> - 2014-12-03 16:43 +0100
  Re: Why compiler is able to vectorize on int but not on unsigned red floyd <no.spam@its.invalid> - 2014-12-03 08:59 -0800
  Re: Why compiler is able to vectorize on int but not on unsigned Alain Ketterlin <alain@dpt-info.u-strasbg.fr> - 2014-12-03 18:53 +0100
    Re: Why compiler is able to vectorize on int but not on unsigned James Kuyper <jameskuyper@verizon.net> - 2014-12-03 13:13 -0500
      Re: Why compiler is able to vectorize on int but not on unsigned Alain Ketterlin <alain@dpt-info.u-strasbg.fr> - 2014-12-04 22:04 +0100
        Re: Why compiler is able to vectorize on int but not on unsigned Melzzzzz <mel@zzzzz.com> - 2014-12-05 10:07 +0100
          Re: Why compiler is able to vectorize on int but not on unsigned Alain Ketterlin <alain@dpt-info.u-strasbg.fr> - 2014-12-05 10:42 +0100
            Re: Why compiler is able to vectorize on int but not on unsigned Melzzzzz <mel@zzzzz.com> - 2014-12-05 11:29 +0100

csiph-web