Page MenuHomePhabricator

wash (Bryce Adelstein Lelbach)Email Not Verified
User

Projects

User does not belong to any projects.

User Details

User Since
Sep 27 2016, 10:32 PM (196 w, 3 d)

Recent Activity

Jan 2 2020

wash created D72106: Remove unnecessary typenames from std/numerics/c.math/abs.pass.cpp.
Jan 2 2020, 1:50 PM · Restricted Project

Aug 8 2019

wash created D65960: Fix environment variable passing in libc++'s `SSHExecutor` `lit` utility.
Aug 8 2019, 10:21 AM · Restricted Project, Restricted Project

Jun 4 2019

wash added a comment to D62874: Fix libc++'s lit configuration language dialect inference for GCC 5 and GCC 6.

I'm okay with this (I don't care much about GCC), but I want to point out you're basically reducing the usefulness of the test suite on GCC 5/6 by just wholesale removing all C++17 tests, when GCC5/6 can actually run a large subset of the tests.

Jun 4 2019, 1:25 PM · Restricted Project
wash created D62874: Fix libc++'s lit configuration language dialect inference for GCC 5 and GCC 6.
Jun 4 2019, 12:59 PM · Restricted Project

Jun 9 2017

wash added a comment to D34038: Implement the non-parallel versions of exclusive_scan and transform_exclusive_scan.

The init typedef test looks good to me.

Jun 9 2017, 7:02 PM

Jun 8 2017

wash added a comment to D34038: Implement the non-parallel versions of exclusive_scan and transform_exclusive_scan.

This implementation works, but performs unnecessary operations.

Jun 8 2017, 4:53 PM
wash added a comment to D34038: Implement the non-parallel versions of exclusive_scan and transform_exclusive_scan.

https://gist.github.com/brycelelbach/137f1e45b737d615134e228ec0c84f3a <- some crude tests I wrote based on slides/notes of mine.

Jun 8 2017, 4:29 PM
wash added inline comments to D34038: Implement the non-parallel versions of exclusive_scan and transform_exclusive_scan.
Jun 8 2017, 4:02 PM
wash added a comment to D33997: Implement the non-execution policy versions of `reduce` and `transform_reduce` for C++17.

I think we need a test case like this for all of the transform_*s

Jun 8 2017, 3:51 PM
wash added a comment to D34007: Implement inclusive_scan and transform_inclusive_scan.
template <class _InputIterator, class _OutputIterator, class _BinaryOperation>
inline _LIBCPP_INLINE_VISIBILITY
_OutputIterator
partial_sum(_InputIterator __first, _InputIterator __last, _OutputIterator __result,
            _BinaryOperation __binary_op)
{
    if (__first != __last)
    {
        typename iterator_traits<_InputIterator>::value_type __t(*__first);
        *__result = __t;
        for (++__first, (void) ++__result; __first != __last; ++__first, (void) ++__result)
        {
            __t = __binary_op(__t, *__first);
            *__result = __t;
        }
    }
    return __result;
}
Jun 8 2017, 3:45 PM
wash added a comment to D34007: Implement inclusive_scan and transform_inclusive_scan.

So, the inclusive_scan overloads that do not take an init parameter should be equivalent to partial_sum for the non-parallel version.

Jun 8 2017, 3:35 PM
wash added a comment to D34007: Implement inclusive_scan and transform_inclusive_scan.

Minor note: there's a mix of tabs and spaces in this diff.

Jun 8 2017, 3:33 PM
wash added a comment to D33997: Implement the non-execution policy versions of `reduce` and `transform_reduce` for C++17.

For all the transform_ variants, the spec allows the "inner" operation (the transformation)'s return type is not constrained. We should have a test for this.

Jun 8 2017, 3:06 PM
wash added a comment to D33997: Implement the non-execution policy versions of `reduce` and `transform_reduce` for C++17.

I think the test reduce_iter_iter_T.pass.cpp can be improved a little bit.

Jun 8 2017, 12:50 PM
wash added a comment to D33997: Implement the non-execution policy versions of `reduce` and `transform_reduce` for C++17.

Suppose you have:

Jun 8 2017, 12:36 PM