On Wed, Aug 13, 2014 at 3:36 PM, Bin.Cheng <amker.ch...@gmail.com> wrote:
> On Sun, Aug 10, 2014 at 2:00 AM, Ulrich Drepper <drep...@gmail.com> wrote:
>> On Sat, Aug 9, 2014 at 1:40 PM, Marc Glisse <marc.gli...@inria.fr> wrote:
>>> __x = result_type(2.0) * __aurng() - 1.0;
>>
>> You're right, we of course need the negatives as well.
>>
>>> Assuming the 2 coordinates are obtained through a rescaling x->2*x-1, if
>>> __sq is not exactly 0, it must be between 2^-103 and 1 (for ieee
>>> double), so I am not sure hypot gains that much (at least in my mind
>>> hypot was mostly a gain close to 0 or infinity, but maybe it has more
>>> advantages). It can only hurt speed though, so not a big issue.
>>
>> Depending on how similar in size the two values are, not using hypot()
>> can drop quite a few bits. Especially with the scaling through
>> division this error can be noticeable. Better be sure. Maybe at some
>> point I have time to investigate the worst case scenario for the
>> numbers in question but until this shows hypot isn't needed it's best
>> to leave it in.
>>
>> I've committed the patch.
>
> Hi,
>
> This causes lots of tests under libstdc++-v3/testsuite/ext/ failed on
> aarch64-none-elf and arm-none-eabi with below log message.
>
> In file included from
> /home/binche01/work/oban-work/build-aarch64-none-elf/obj/gcc2/aarch64-none-elf/libstdc++-v3/include/ext/random:3494:0,
> from
> /home/binche01/work/oban-work/src/gcc/libstdc++-v3/testsuite/ext/random/arcsine_distribution/cons/default.cc:23:
> /home/binche01/work/oban-work/src/gcc/libstdc++-v3/include/ext/random.tcc:
> In member function
> '__gnu_cxx::{anonymous}::uniform_on_sphere_helper<2ul,
> _RealType>::result_type
> __gnu_cxx::{anonymous}::uniform_on_sphere_helper<2ul,
> _RealType>::operator()(_NormalDistribution&,
> _UniformRandomNumberGenerator&)':
> /home/binche01/work/oban-work/src/gcc/libstdc++-v3/include/ext/random.tcc:1609:18:
> error: 'hypot' is not a member of 'std'
> auto __norm = std::hypot(__ret[0], __ret[1]);
> ^
> /home/binche01/work/oban-work/src/gcc/libstdc++-v3/include/ext/random.tcc:1609:18:
> note: suggested alternative:
> In file included from
> /home/binche01/work/oban-work/build-aarch64-none-elf/obj/gcc2/aarch64-none-elf/libstdc++-v3/include/cmath:44:0,
> from
> /home/binche01/work/oban-work/build-aarch64-none-elf/obj/gcc2/aarch64-none-elf/libstdc++-v3/include/random:38,
> from
> /home/binche01/work/oban-work/build-aarch64-none-elf/obj/gcc2/aarch64-none-elf/libstdc++-v3/include/ext/random:38,
> from
> /home/binche01/work/oban-work/src/gcc/libstdc++-v3/testsuite/ext/random/arcsine_distribution/cons/default.cc:23:
> /home/binche01/work/oban-work/target-aarch64-none-elf/aarch64-none-elf/include/math.h:306:15:
> note: 'hypot'
> extern double hypot _PARAMS((double, double));
> ^
>
> FAIL: ext/random/arcsine_distribution/cons/default.cc (test for excess errors)
> Excess errors:
> /home/binche01/work/oban-work/src/gcc/libstdc++-v3/include/ext/random.tcc:1609:18:
> error: 'hypot' is not a member of 'std'
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62118 filed.
Thanks,
bin
