Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix omp vectorization pragma errors #6553

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
+69 −71
Open

Fix omp vectorization pragma errors #6553

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9484458
Fix omp vectorization pragma errors
Pansysk75 Oct 10, 2024
90b21d4
Silence simd warnings on CI
Pansysk75 Oct 16, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions .jenkins/lsu/env-clang-17.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,7 @@ module load openmpi
export HPXRUN_RUNWRAPPER=srun
export CXX_STD="20"
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:$PKG_CONFIG_PATH
export CXXFLAGS="-fopenmp -Wno-pass-failed ${CXXFLAGS}"

configure_extra_options+=" -DCMAKE_BUILD_TYPE=${build_type}"
configure_extra_options+=" -DHPX_WITH_CXX_STANDARD=${CXX_STD}"
Expand Down
1 change: 1 addition & 0 deletions .jenkins/lsu/env-gcc-13.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,7 @@ module load openmpi
export HPXRUN_RUNWRAPPER=srun
export CXX_STD="20"
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:$PKG_CONFIG_PATH
export CXXFLAGS="-fopenmp ${CXXFLAGS}"

configure_extra_options+=" -DHPX_WITH_CXX_STANDARD=${CXX_STD}"
configure_extra_options+=" -DHPX_WITH_MALLOC=system"
Expand Down
39 changes: 18 additions & 21 deletions libs/core/algorithms/include/hpx/parallel/unseq/loop.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@ namespace hpx::parallel::util {
struct unseq_loop_n
{
template <typename InIter, typename F>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr InIter call(
HPX_HOST_DEVICE HPX_FORCEINLINE static InIter call(
InIter HPX_RESTRICT it, std::size_t num, F&& f)
{
// clang-format off
Expand Down Expand Up @@ -135,10 +135,9 @@ namespace hpx::parallel::util {
struct unseq_loop2
{
template <typename InIter1, typename InIter2, typename F>
HPX_HOST_DEVICE
HPX_FORCEINLINE static constexpr std::pair<InIter1, InIter2>
call(InIter1 HPX_RESTRICT it1, InIter1 HPX_RESTRICT last1,
InIter2 HPX_RESTRICT it2, F&& f)
HPX_HOST_DEVICE HPX_FORCEINLINE static std::pair<InIter1, InIter2>
call(InIter1 HPX_RESTRICT it1, InIter1 HPX_RESTRICT last1,
InIter2 HPX_RESTRICT it2, F&& f)
{
constexpr bool iterators_are_random_access =
hpx::traits::is_random_access_iterator_v<InIter1> &&
Expand Down Expand Up @@ -216,19 +215,18 @@ namespace hpx::parallel::util {

// clang-format off
HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (std::size_t i = 0; i < count;
(void) ++it, i += 4) // -V112
for (std::size_t i = 0; i < count; i += 4) // -V112
{
HPX_INVOKE(f, it);
HPX_INVOKE(f, ++it);
HPX_INVOKE(f, ++it);
HPX_INVOKE(f, ++it);
HPX_INVOKE(f, it++);
HPX_INVOKE(f, it++);
HPX_INVOKE(f, it++);
HPX_INVOKE(f, it++);
}

HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (/**/; count < num; (void) ++count, ++it)
for (std::size_t i = count; i < num; ++i)
{
HPX_INVOKE(f, it);
HPX_INVOKE(f, it++);
}
// clang-format on

Expand Down Expand Up @@ -256,19 +254,18 @@ namespace hpx::parallel::util {

// clang-format off
HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (std::size_t i = 0; i < count; (void) ++it, ++dest,
i += 4) // -V112
for (std::size_t i = 0; i < count; i += 4) // -V112
{
HPX_INVOKE(f, it, dest);
HPX_INVOKE(f, ++it, ++dest);
HPX_INVOKE(f, ++it, ++dest);
HPX_INVOKE(f, ++it, ++dest);
HPX_INVOKE(f, it++, dest++);
HPX_INVOKE(f, it++, dest++);
HPX_INVOKE(f, it++, dest++);
HPX_INVOKE(f, it++, dest++);
}

HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (/**/; count < num; (void) ++count, ++it, ++dest)
for (std::size_t i = count; i < num; ++i)
{
HPX_INVOKE(f, it, dest);
HPX_INVOKE(f, it++, dest++);
}
//clang-format on

Expand Down
34 changes: 17 additions & 17 deletions libs/core/algorithms/include/hpx/parallel/unseq/reduce_helpers.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -85,7 +85,7 @@ namespace hpx::parallel::util::detail {
{
#if defined(HPX_HAVE_VECTOR_REDUCTION)
template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_plus_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce /* */, Convert conv)
{
Expand All @@ -99,7 +99,7 @@ namespace hpx::parallel::util::detail {
}

template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_minus_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce /* */, Convert conv)
{
Expand All @@ -113,7 +113,7 @@ namespace hpx::parallel::util::detail {
}

template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_multiplies_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce /* */, Convert conv)
{
Expand All @@ -127,7 +127,7 @@ namespace hpx::parallel::util::detail {
}

template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_bit_and_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce /* */, Convert conv)
{
Expand All @@ -141,7 +141,7 @@ namespace hpx::parallel::util::detail {
}

template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_bit_or_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce /* */, Convert conv)
{
Expand All @@ -155,7 +155,7 @@ namespace hpx::parallel::util::detail {
}

template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_bit_xor_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce /* */, Convert conv)
{
Expand All @@ -169,7 +169,7 @@ namespace hpx::parallel::util::detail {
}

template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_logical_and_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce /* */, Convert conv)
{
Expand All @@ -183,7 +183,7 @@ namespace hpx::parallel::util::detail {
}

template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_logical_or_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce /* */, Convert conv)
{
Expand All @@ -197,7 +197,7 @@ namespace hpx::parallel::util::detail {
}
#endif
template <typename Iter1, typename T, typename Convert, typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_not_omp_reduction_v<T, Reduce>, T>
reduce(Iter1 it, std::size_t count, T init, Reduce r, Convert conv)
{
Expand Down Expand Up @@ -274,7 +274,7 @@ namespace hpx::parallel::util::detail {
#if defined(HPX_HAVE_VECTOR_REDUCTION)
template <typename Iter1, typename Iter2, typename T, typename Convert,
typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_plus_reduction_v<T, Reduce>, T>
reduce(Iter1 it1, Iter2 it2, std::size_t count, T init, Reduce /* */,
Convert conv)
Expand All @@ -290,7 +290,7 @@ namespace hpx::parallel::util::detail {

template <typename Iter1, typename Iter2, typename T, typename Convert,
typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_minus_reduction_v<T, Reduce>, T>
reduce(Iter1 it1, Iter2 it2, std::size_t count, T init, Reduce /* */,
Convert conv)
Expand All @@ -306,7 +306,7 @@ namespace hpx::parallel::util::detail {

template <typename Iter1, typename Iter2, typename T, typename Convert,
typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_multiplies_reduction_v<T, Reduce>, T>
reduce(Iter1 it1, Iter2 it2, std::size_t count, T init, Reduce /* */,
Convert conv)
Expand All @@ -322,7 +322,7 @@ namespace hpx::parallel::util::detail {

template <typename Iter1, typename Iter2, typename T, typename Convert,
typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_bit_and_reduction_v<T, Reduce>, T>
reduce(Iter1 it1, Iter2 it2, std::size_t count, T init, Reduce /* */,
Convert conv)
Expand All @@ -338,7 +338,7 @@ namespace hpx::parallel::util::detail {

template <typename Iter1, typename Iter2, typename T, typename Convert,
typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_bit_or_reduction_v<T, Reduce>, T>
reduce(Iter1 it1, Iter2 it2, std::size_t count, T init, Reduce /* */,
Convert conv)
Expand Down Expand Up @@ -370,7 +370,7 @@ namespace hpx::parallel::util::detail {

template <typename Iter1, typename Iter2, typename T, typename Convert,
typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_logical_and_reduction_v<T, Reduce>, T>
reduce(Iter1 it1, Iter2 it2, std::size_t count, T init, Reduce /* */,
Convert conv)
Expand All @@ -386,7 +386,7 @@ namespace hpx::parallel::util::detail {

template <typename Iter1, typename Iter2, typename T, typename Convert,
typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_arithmetic_logical_or_reduction_v<T, Reduce>, T>
reduce(Iter1 it1, Iter2 it2, std::size_t count, T init, Reduce /* */,
Convert conv)
Expand All @@ -402,7 +402,7 @@ namespace hpx::parallel::util::detail {
#endif
template <typename Iter1, typename Iter2, typename T, typename Convert,
typename Reduce>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr std::enable_if_t<
HPX_HOST_DEVICE HPX_FORCEINLINE static std::enable_if_t<
is_not_omp_reduction_v<T, Reduce>, T>
reduce(Iter1 it1, Iter2 it2, std::size_t count, T init, Reduce r,
Convert conv)
Expand Down
32 changes: 16 additions & 16 deletions libs/core/algorithms/include/hpx/parallel/unseq/transform_loop.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,10 +25,9 @@ namespace hpx::parallel::util {
struct unseq_transform_loop_n
{
template <typename InIter, typename OutIter, typename F>
HPX_HOST_DEVICE
HPX_FORCEINLINE static constexpr std::pair<InIter, OutIter>
call(InIter HPX_RESTRICT it, std::size_t num,
OutIter HPX_RESTRICT dest, F&& f)
HPX_HOST_DEVICE HPX_FORCEINLINE static std::pair<InIter, OutIter>
call(InIter HPX_RESTRICT it, std::size_t num,
OutIter HPX_RESTRICT dest, F&& f)
{
constexpr bool iterators_are_random_access =
hpx::traits::is_random_access_iterator_v<InIter> &&
Expand Down Expand Up @@ -75,10 +74,9 @@ namespace hpx::parallel::util {
struct unseq_transform_loop_n_ind
{
template <typename InIter, typename OutIter, typename F>
HPX_HOST_DEVICE
HPX_FORCEINLINE static constexpr std::pair<InIter, OutIter>
call(InIter HPX_RESTRICT it, std::size_t num,
OutIter HPX_RESTRICT dest, F&& f)
HPX_HOST_DEVICE HPX_FORCEINLINE static std::pair<InIter, OutIter>
call(InIter HPX_RESTRICT it, std::size_t num,
OutIter HPX_RESTRICT dest, F&& f)
{
constexpr bool iterators_are_random_access =
hpx::traits::is_random_access_iterator_v<InIter> &&
Expand Down Expand Up @@ -240,10 +238,11 @@ namespace hpx::parallel::util {
{
template <typename InIter1, typename InIter2, typename OutIter,
typename F>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr hpx::tuple<InIter1,
InIter2, OutIter>
call(InIter1 HPX_RESTRICT first1, std::size_t num,
InIter2 HPX_RESTRICT first2, OutIter HPX_RESTRICT dest, F&& f)
HPX_HOST_DEVICE
HPX_FORCEINLINE static hpx::tuple<InIter1, InIter2, OutIter>
call(InIter1 HPX_RESTRICT first1, std::size_t num,
InIter2 HPX_RESTRICT first2, OutIter HPX_RESTRICT dest,
F&& f)
{
constexpr bool iterators_are_random_access =
hpx::traits::is_random_access_iterator_v<InIter1> &&
Expand Down Expand Up @@ -401,10 +400,11 @@ namespace hpx::parallel::util {
{
template <typename InIter1, typename InIter2, typename OutIter,
typename F>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr hpx::tuple<InIter1,
InIter2, OutIter>
call(InIter1 HPX_RESTRICT first1, std::size_t num,
InIter2 HPX_RESTRICT first2, OutIter HPX_RESTRICT dest, F&& f)
HPX_HOST_DEVICE
HPX_FORCEINLINE static hpx::tuple<InIter1, InIter2, OutIter>
call(InIter1 HPX_RESTRICT first1, std::size_t num,
InIter2 HPX_RESTRICT first2, OutIter HPX_RESTRICT dest,
F&& f)
{
constexpr bool iterators_are_random_access =
hpx::traits::is_random_access_iterator_v<InIter1> &&
Expand Down
33 changes: 16 additions & 17 deletions libs/core/algorithms/include/hpx/parallel/util/loop.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,14 +58,14 @@ namespace hpx::parallel::util {
HPX_CONCEPT_REQUIRES_( // forces hpx::execution::unseq
hpx::is_unsequenced_execution_policy_v<ExPolicy> &&
!hpx::is_parallel_execution_policy_v<ExPolicy>)>
HPX_HOST_DEVICE HPX_FORCEINLINE static constexpr Begin call(
HPX_HOST_DEVICE HPX_FORCEINLINE static Begin call(
ExPolicy&&, Begin it, End end, F&& f)
{
// clang-format off
HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (/**/; it != end; ++it)
for (Begin& iter = it; iter != end; ++iter)
{
HPX_INVOKE(f, it);
HPX_INVOKE(f, iter);
}
// clang-format on

Expand Down Expand Up @@ -695,8 +695,9 @@ namespace hpx::parallel::util {
{
// clang-format off
HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (/**/; it != last; (void) ++it)
HPX_INVOKE(f, it);
for (FwdIter& iter = it; iter != last; ++iter){
HPX_INVOKE(f, iter);
}
// clang-format on

return it;
Expand All @@ -721,8 +722,8 @@ namespace hpx::parallel::util {
{
// clang-format off
HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (/**/; it != last; (void) ++it, ++dest)
f(it, dest);
for (Iter& iter = it; iter != last; ++iter)
f(iter, dest++);
// clang-format on

return dest;
Expand Down Expand Up @@ -913,16 +914,15 @@ namespace hpx::parallel::util {

// clang-format off
HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (std::size_t i = 0; i < count;
(void) ++it, ++dest, ++i) // -V112
for (std::size_t i = 0; i < count; ++i) // -V112
{
HPX_INVOKE(f, it, dest);
HPX_INVOKE(f, it++, dest++);
}

HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (/**/; count < num; (void) ++count, ++it, ++dest)
for (std::size_t i = count; i < num; ++i)
{
HPX_INVOKE(f, it, dest);
HPX_INVOKE(f, it++, dest++);
}
// clang-format on

Expand Down Expand Up @@ -950,16 +950,15 @@ namespace hpx::parallel::util {

// clang-format off
HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (std::size_t i = 0; i < count;
(void) ++it, ++i) // -V112
for (std::size_t i = 0; i < count; ++i) // -V112
{
HPX_INVOKE(f, it);
HPX_INVOKE(f, it++);
}

HPX_IVDEP HPX_UNROLL HPX_VECTORIZE
for (/**/; count < num; (void) ++count, ++it)
for (std::size_t i = count; i < num; ++i)
{
HPX_INVOKE(f, it);
HPX_INVOKE(f, it++);
}
// clang-format on

Expand Down
Loading