Re: [boost] Boost.Convert. Take 2.

Subject: Re: [boost] Boost.Convert. Take 2.
From: Alex Hagen-Zanker (a.hagen-zanker_at_[hidden])
Date: 2014-02-22 07:57:14

> -----Original Message-----
> From: Boost [mailto:boost-bounces_at_[hidden]] On Behalf Of Rob
> Stewart
> Sent: 21 February 2014 19:29
> To: boost_at_[hidden]
> Subject: Re: [boost] Boost.Convert. Take 2.
> 
> On February 21, 2014 12:21:12 PM EST, alex <alexhighviz_at_[hidden]>
> wrote:
> >
> >
> >>-----Original Message-----
> >>From: Boost [mailto:boost-bounces_at_[hidden]] On Behalf Of Rob
> >Stewart
> >>Sent: 21 February 2014 15:35
> >>To: boost_at_[hidden]
> >>Subject: Re: [boost] Boost.Convert. Take 2.
> >>
> >>On February 21, 2014 5:48:32 AM EST, alex <alexhighviz_at_[hidden]>
> >>wrote:
> >>>
> >>>
> >>>>>All loops are actually like
> >>>>>
> >>>>>for (int k = 0; k < local::num_cycles; ++k) {
> >>>>>    int k = boost::convert<int>::from("12345", ccnv).value();
> >>>>>    BOOST_ASSERT(k == 12345);
> >>>>>}
> >>>>>
> >>>>>Do you think there still might be something I missed?
> >>>>
> >>>>The assertion won't help in an optimized build, of course. Save the
> >>>values and
> >>>>print them at the end,  after you've captured the elapsed times.
> >>>>
> >>>
> >>>Isn't it also a problem that the inside of the loop doesn't vary? A
> >>>smart compiler may decide to evaluate
> >>>boost::convert<int>::from("12345",
> >>>ccnv).value() only once and re-use the result?
> >>
> >>Of course,  yes. I should have noticed that. Incrementing a value each
> >iteration
> >>and printing that value at the end should work. Overflow isn't
> >important.
> >Just
> >>using the final value in an opaque way will prevent optimizing away
> >the
> >>conversions in the loops.
> >>
> >>
> >That seems easier said than done, because it is the value of a string
> >that needs to be unpredictable?
> >
> >int sum  = 0;
> >for (int k = 0; k < local::num_cycles; ++k) {
> >    char str[] = "12345";
> >    str[4 - k % 5] = 49 + k % 9; //because the char '1' has value 49
> >    sum += boost::convert<int>::from(str, ccnv).value(); } std::cout <<
> >sum;
> 
> I'm not sure that's necessary. Unless there is a lot more inlining than
I'm
> imagining,  I don't think the optimizer will cause a problem with the
string
> literal input.
> 
I'm not sure either of course. But if I were a compiler, I would not
repeatedly do the same computation on two constants ("12345" and ccnv]. I am
not sure however if ccnv's mutable stream_  member will make the compiler
dismiss the const status of ccnv.