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

Use __builtin_shufflevector for the generic implementation of shuffle #958

Merged
merged 1 commit into from
Oct 18, 2023
Merged

Use __builtin_shufflevector for the generic implementation of shuffle #958

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
17 changes: 17 additions & 0 deletions include/xsimd/arch/generic/xsimd_generic_memory.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -419,12 +419,29 @@ namespace xsimd
return select(batch_bool_constant<batch<T, A>, (Indices < bsize)...>(), x, y);
}

#if defined(__has_builtin)
#if __has_builtin(__builtin_shuffle_vector)
#define builtin_shuffle __builtin_shuffle_vector
#endif
#endif

#if defined(builtin_shuffle)
return builtin_shuffle(x.data, y.data, Indices...);

// FIXME: my experiments show that GCC only correctly optimizes this builtin
// starting at GCC 13, where it already has __builtin_shuffle_vector
//
// #elif __has_builtin(__builtin_shuffle) || GCC >= 6
// typedef ITy integer_vector_type __attribute__((vector_size(sizeof(batch<ITy, A>))));
// return __builtin_shuffle(x.data, y.data, integer_vector_type{Indices...});
#else
// Use a generic_pattern. It is suboptimal but clang optimizes this
// pretty well.
batch<T, A> x_lane = swizzle(x, batch_constant<batch<ITy, A>, ((Indices >= bsize) ? (Indices - bsize) : Indices)...>());
batch<T, A> y_lane = swizzle(y, batch_constant<batch<ITy, A>, ((Indices >= bsize) ? (Indices - bsize) : Indices)...>());
batch_bool_constant<batch<T, A>, (Indices < bsize)...> select_x_lane;
return select(select_x_lane, x_lane, y_lane);
#endif
}

// store
Expand Down