shamalgs::primitives Namespace Reference

namespace for primitive algorithm (e.g. sort, scan, reductions, ...) More...

Namespaces
namespace	impl
	namespace to control implementation behavior

Classes
class	ImplControl

Enumerations
enum class	IS_ALL_TRUE_IMPL : u32 { HOST , SUM_REDUCTION }
enum class	REDUCTION_IMPL : u32 { FALLBACK }
enum class	EXSCAN_IN_PLACE_IMPL : u32 { STD_SCAN }

Functions
template<class T >
void	append_subset_to (const sham::DeviceBuffer< T > &buf, const sham::DeviceBuffer< u32 > &idxs_buf, u32 nvar, sham::DeviceBuffer< T > &buf_other, u32 start_enque)
	Appends a subset of elements from one buffer to another.
template<class Tkey >
constexpr void	binary_range_search (const Tkey *__restrict__ key, u32 first, u32 last, const Tkey &value_min, const Tkey &value_max, u32 &inf, u32 &sup)
	Find the range of indices for which key[inf] <= value_min <= value_max <= key[sup].
template<class T , class Tbins , class... Targs, class Tfunctor >
sham::DeviceBuffer< T >	compute_histogram (const sham::DeviceScheduler_ptr &dev_sched, const sham::DeviceBuffer< Tbins > &bin_edge_inf, const sham::DeviceBuffer< Tbins > &bin_edge_sup, size_t element_count, Tfunctor &&functor, const sham::DeviceBuffer< Targs > &...input_data)
template<class T >
sham::DeviceBuffer< T >	compute_histogram_basic (const sham::DeviceScheduler_ptr &dev_sched, const sham::DeviceBuffer< T > &bin_edge_inf, const sham::DeviceBuffer< T > &bin_edge_sup, const sham::DeviceBuffer< T > &positions)
template<class T >
shambase::VecComponent< T >	dot_sum (sham::DeviceBuffer< T > &buf1, u32 start_id, u32 end_id)
	Compute the sum of dot products of elements in a device buffer with themselves.
template<class T >
shambase::VecComponent< T >	dot_sum (sham::DeviceBuffer< T > &buf1)
template<class T >
bool	equals (sycl::queue &q, sycl::buffer< T > &buf1, sycl::buffer< T > &buf2, u32 cnt)
	Compare elements between two sycl::buffers for equality.
template<class T >
bool	equals (const sham::DeviceScheduler_ptr &dev_sched, sham::DeviceBuffer< T > &buf1, sham::DeviceBuffer< T > &buf2, u32 cnt)
	Compare elements between two sham::DeviceBuffers for equality.
template<class T >
bool	equals (const sham::DeviceScheduler_ptr &q, sham::DeviceBuffer< T > &buf1, sham::DeviceBuffer< T > &buf2)
	Compare all elements between two sham::DeviceBuffers for equality.
template<class T >
bool	equals (sycl::queue &q, sycl::buffer< T > &buf1, sycl::buffer< T > &buf2)
	Compare all elements between two sycl::buffers for equality.
template<class T >
bool	equals_ptr_s (sycl::queue &q, const std::unique_ptr< sycl::buffer< T > > &buf1, const std::unique_ptr< sycl::buffer< T > > &buf2, u32 cnt)
	Compare elements between two unique_ptr-wrapped sycl::buffers with count.
template<class T >
bool	equals_ptr (sycl::queue &q, const std::unique_ptr< sycl::buffer< T > > &buf1, const std::unique_ptr< sycl::buffer< T > > &buf2)
	Compare all elements between two unique_ptr-wrapped sycl::buffers.
template<class Tvec , sham::USMKindTarget target>
sham::DeviceBuffer< typename shambase::VectorProperties< Tvec >::component_type, target >	flatten_buffer (const sham::DeviceBuffer< Tvec, target > &buffer)
	Flatten a buffer of vector type into a buffer of scalar type.
template<class Tvec , sham::USMKindTarget target>
sham::DeviceBuffer< Tvec, target >	unflatten_buffer (const sham::DeviceBuffer< typename shambase::VectorProperties< Tvec >::component_type, target > &buffer)
	Unflatten a buffer that contains a flattened vector.
sham::DeviceBuffer< u32 >	gen_buffer_index (sham::DeviceScheduler_ptr sched, u32 len)
	Generates a buffer where buf[i] = i.
void	fill_buffer_index (sham::DeviceBuffer< u32 > &buf, u32 len)
	Fills a buffer with sequential indices, such that buf[i] = i.
template<class T >
bool	is_all_true (sycl::buffer< T > &buf, u32 cnt)
	Check if all elements in a sycl::buffer are non-zero.
template<class T >
bool	is_all_true (sham::DeviceBuffer< T > &buf, u32 cnt)
	Check if all elements in a sham::DeviceBuffer are non-zero.
template<typename Tval >
sham::DeviceBuffer< Tval >	linspace (Tval Rmin, Tval Rmax, u32 N)
	Create an array of N values between two values.
template<class Tkey >
constexpr u32	binary_search_lower_bound (const Tkey *__restrict__ key, u32 first, u32 last, const Tkey &value)
	GPU compatible implementation of std::lower_bound.
template<class T , class Engine = std::mt19937>
T	mock_value (Engine &eng, T min_bound, T max_bound)
	Generates a random mock value within specified bounds.
template<class T , class Engine = std::mt19937>
T	mock_value (Engine &eng)
	Generates a random mock value using default bounds.
template<class T >
std::vector< T >	mock_vector (u64 seed, u32 len, T min_bound, T max_bound)
	Generates a vector of random mock values within specified bounds.
template<class T >
std::vector< T >	mock_vector (u64 seed, u32 len)
	Generates a vector of random mock values using default bounds.
template<class T >
T	sum (const sham::DeviceScheduler_ptr &sched, const sham::DeviceBuffer< T > &buf1, u32 start_id, u32 end_id)
	Compute the sum of elements in a device buffer within a specified range.
template<class T >
T	min (const sham::DeviceScheduler_ptr &sched, const sham::DeviceBuffer< T > &buf1, u32 start_id, u32 end_id)
	Find the minimum element in a device buffer within a specified range.
template<class T >
T	max (const sham::DeviceScheduler_ptr &sched, const sham::DeviceBuffer< T > &buf1, u32 start_id, u32 end_id)
	Find the maximum element in a device buffer within a specified range.
template<class T >
void	scan_exclusive_sum_in_place (sham::DeviceBuffer< T > &buf1, u32 len)
	Compute exclusive prefix sum in-place on a device buffer.
template<class T >
void	segmented_sort_in_place (sham::DeviceBuffer< T > &buf, const sham::DeviceBuffer< u32 > &offsets)
template<class Tkey , class Tval >
void	sort_by_key_pow2_len (sycl::queue &q, sycl::buffer< Tkey > &buf_key, sycl::buffer< Tval > &buf_values, u32 len)
	Sort key-value pairs using sycl::buffers (power-of-2 optimized)
template<class Tkey , class Tval >
void	sort_by_key_pow2_len (const sham::DeviceScheduler_ptr &sched, sham::DeviceBuffer< Tkey > &buf_key, sham::DeviceBuffer< Tval > &buf_values, u32 len)
	Sort key-value pairs using USM buffers (power-of-2 optimized)
template<class Tkey , class Tval >
void	sort_by_key (sycl::queue &q, sycl::buffer< Tkey > &buf_key, sycl::buffer< Tval > &buf_values, u32 len)
	Sort key-value pairs using sycl::buffers.
template<class Tkey , class Tval >
void	sort_by_key (const sham::DeviceScheduler_ptr &sched, sham::DeviceBuffer< Tkey > &buf_key, sham::DeviceBuffer< Tval > &buf_values, u32 len)
	Sort key-value pairs using USM buffers.
template<class Tkey >
constexpr u32	binary_search_upper_bound (const Tkey *__restrict__ key, u32 first, u32 last, const Tkey &value)
	GPU compatible implementation of std::upper_bound.
IS_ALL_TRUE_IMPL	is_all_true_impl_from_params (const std::string &impl)
shamalgs::impl_param	is_all_true_impl_to_params (const IS_ALL_TRUE_IMPL &impl)
template bool	is_all_true (sham::DeviceBuffer< u8 > &buf, u32 cnt)
template<>
u8	mock_value (std::mt19937 &eng, u8 min_bound, u8 max_bound)
template<>
u16	mock_value (std::mt19937 &eng, u16 min_bound, u16 max_bound)
template<>
u32	mock_value (std::mt19937 &eng, u32 min_bound, u32 max_bound)
template<>
u64	mock_value (std::mt19937 &eng, u64 min_bound, u64 max_bound)
template<>
i8	mock_value (std::mt19937 &eng, i8 min_bound, i8 max_bound)
template<>
i16	mock_value (std::mt19937 &eng, i16 min_bound, i16 max_bound)
template<>
i32	mock_value (std::mt19937 &eng, i32 min_bound, i32 max_bound)
template<>
i64	mock_value (std::mt19937 &eng, i64 min_bound, i64 max_bound)
template<>
f32	mock_value (std::mt19937 &eng, f32 min_bound, f32 max_bound)
template<>
f64	mock_value (std::mt19937 &eng, f64 min_bound, f64 max_bound)
template<>
sycl::vec< u64, 2 >	mock_value (std::mt19937 &eng, sycl::vec< u64, 2 > min_bound, sycl::vec< u64, 2 > max_bound)
template<>
sycl::vec< u64, 3 >	mock_value (std::mt19937 &eng, sycl::vec< u64, 3 > min_bound, sycl::vec< u64, 3 > max_bound)
template<>
sycl::vec< u64, 4 >	mock_value (std::mt19937 &eng, sycl::vec< u64, 4 > min_bound, sycl::vec< u64, 4 > max_bound)
template<>
sycl::vec< u64, 8 >	mock_value (std::mt19937 &eng, sycl::vec< u64, 8 > min_bound, sycl::vec< u64, 8 > max_bound)
template<>
sycl::vec< u64, 16 >	mock_value (std::mt19937 &eng, sycl::vec< u64, 16 > min_bound, sycl::vec< u64, 16 > max_bound)
template<>
sycl::vec< u32, 2 >	mock_value (std::mt19937 &eng, sycl::vec< u32, 2 > min_bound, sycl::vec< u32, 2 > max_bound)
template<>
sycl::vec< u32, 3 >	mock_value (std::mt19937 &eng, sycl::vec< u32, 3 > min_bound, sycl::vec< u32, 3 > max_bound)
template<>
sycl::vec< u32, 4 >	mock_value (std::mt19937 &eng, sycl::vec< u32, 4 > min_bound, sycl::vec< u32, 4 > max_bound)
template<>
sycl::vec< u32, 8 >	mock_value (std::mt19937 &eng, sycl::vec< u32, 8 > min_bound, sycl::vec< u32, 8 > max_bound)
template<>
sycl::vec< u32, 16 >	mock_value (std::mt19937 &eng, sycl::vec< u32, 16 > min_bound, sycl::vec< u32, 16 > max_bound)
template<>
sycl::vec< u16, 2 >	mock_value (std::mt19937 &eng, sycl::vec< u16, 2 > min_bound, sycl::vec< u16, 2 > max_bound)
template<>
sycl::vec< u16, 3 >	mock_value (std::mt19937 &eng, sycl::vec< u16, 3 > min_bound, sycl::vec< u16, 3 > max_bound)
template<>
sycl::vec< u16, 4 >	mock_value (std::mt19937 &eng, sycl::vec< u16, 4 > min_bound, sycl::vec< u16, 4 > max_bound)
template<>
sycl::vec< u16, 8 >	mock_value (std::mt19937 &eng, sycl::vec< u16, 8 > min_bound, sycl::vec< u16, 8 > max_bound)
template<>
sycl::vec< u16, 16 >	mock_value (std::mt19937 &eng, sycl::vec< u16, 16 > min_bound, sycl::vec< u16, 16 > max_bound)
template<>
sycl::vec< u8, 2 >	mock_value (std::mt19937 &eng, sycl::vec< u8, 2 > min_bound, sycl::vec< u8, 2 > max_bound)
template<>
sycl::vec< u8, 3 >	mock_value (std::mt19937 &eng, sycl::vec< u8, 3 > min_bound, sycl::vec< u8, 3 > max_bound)
template<>
sycl::vec< u8, 4 >	mock_value (std::mt19937 &eng, sycl::vec< u8, 4 > min_bound, sycl::vec< u8, 4 > max_bound)
template<>
sycl::vec< u8, 8 >	mock_value (std::mt19937 &eng, sycl::vec< u8, 8 > min_bound, sycl::vec< u8, 8 > max_bound)
template<>
sycl::vec< u8, 16 >	mock_value (std::mt19937 &eng, sycl::vec< u8, 16 > min_bound, sycl::vec< u8, 16 > max_bound)
template<>
sycl::vec< i64, 2 >	mock_value (std::mt19937 &eng, sycl::vec< i64, 2 > min_bound, sycl::vec< i64, 2 > max_bound)
template<>
sycl::vec< i64, 3 >	mock_value (std::mt19937 &eng, sycl::vec< i64, 3 > min_bound, sycl::vec< i64, 3 > max_bound)
template<>
sycl::vec< i64, 4 >	mock_value (std::mt19937 &eng, sycl::vec< i64, 4 > min_bound, sycl::vec< i64, 4 > max_bound)
template<>
sycl::vec< i64, 8 >	mock_value (std::mt19937 &eng, sycl::vec< i64, 8 > min_bound, sycl::vec< i64, 8 > max_bound)
template<>
sycl::vec< i64, 16 >	mock_value (std::mt19937 &eng, sycl::vec< i64, 16 > min_bound, sycl::vec< i64, 16 > max_bound)
template<>
sycl::vec< i32, 2 >	mock_value (std::mt19937 &eng, sycl::vec< i32, 2 > min_bound, sycl::vec< i32, 2 > max_bound)
template<>
sycl::vec< i32, 3 >	mock_value (std::mt19937 &eng, sycl::vec< i32, 3 > min_bound, sycl::vec< i32, 3 > max_bound)
template<>
sycl::vec< i32, 4 >	mock_value (std::mt19937 &eng, sycl::vec< i32, 4 > min_bound, sycl::vec< i32, 4 > max_bound)
template<>
sycl::vec< i32, 8 >	mock_value (std::mt19937 &eng, sycl::vec< i32, 8 > min_bound, sycl::vec< i32, 8 > max_bound)
template<>
sycl::vec< i32, 16 >	mock_value (std::mt19937 &eng, sycl::vec< i32, 16 > min_bound, sycl::vec< i32, 16 > max_bound)
template<>
sycl::vec< i16, 2 >	mock_value (std::mt19937 &eng, sycl::vec< i16, 2 > min_bound, sycl::vec< i16, 2 > max_bound)
template<>
sycl::vec< i16, 3 >	mock_value (std::mt19937 &eng, sycl::vec< i16, 3 > min_bound, sycl::vec< i16, 3 > max_bound)
template<>
sycl::vec< i16, 4 >	mock_value (std::mt19937 &eng, sycl::vec< i16, 4 > min_bound, sycl::vec< i16, 4 > max_bound)
template<>
sycl::vec< i16, 8 >	mock_value (std::mt19937 &eng, sycl::vec< i16, 8 > min_bound, sycl::vec< i16, 8 > max_bound)
template<>
sycl::vec< i16, 16 >	mock_value (std::mt19937 &eng, sycl::vec< i16, 16 > min_bound, sycl::vec< i16, 16 > max_bound)
template<>
sycl::vec< i8, 2 >	mock_value (std::mt19937 &eng, sycl::vec< i8, 2 > min_bound, sycl::vec< i8, 2 > max_bound)
template<>
sycl::vec< i8, 3 >	mock_value (std::mt19937 &eng, sycl::vec< i8, 3 > min_bound, sycl::vec< i8, 3 > max_bound)
template<>
sycl::vec< i8, 4 >	mock_value (std::mt19937 &eng, sycl::vec< i8, 4 > min_bound, sycl::vec< i8, 4 > max_bound)
template<>
sycl::vec< i8, 8 >	mock_value (std::mt19937 &eng, sycl::vec< i8, 8 > min_bound, sycl::vec< i8, 8 > max_bound)
template<>
sycl::vec< i8, 16 >	mock_value (std::mt19937 &eng, sycl::vec< i8, 16 > min_bound, sycl::vec< i8, 16 > max_bound)
template<>
sycl::vec< f64, 2 >	mock_value (std::mt19937 &eng, sycl::vec< f64, 2 > min_bound, sycl::vec< f64, 2 > max_bound)
template<>
sycl::vec< f64, 3 >	mock_value (std::mt19937 &eng, sycl::vec< f64, 3 > min_bound, sycl::vec< f64, 3 > max_bound)
template<>
sycl::vec< f64, 4 >	mock_value (std::mt19937 &eng, sycl::vec< f64, 4 > min_bound, sycl::vec< f64, 4 > max_bound)
template<>
sycl::vec< f64, 8 >	mock_value (std::mt19937 &eng, sycl::vec< f64, 8 > min_bound, sycl::vec< f64, 8 > max_bound)
template<>
sycl::vec< f64, 16 >	mock_value (std::mt19937 &eng, sycl::vec< f64, 16 > min_bound, sycl::vec< f64, 16 > max_bound)
template<>
sycl::vec< f32, 2 >	mock_value (std::mt19937 &eng, sycl::vec< f32, 2 > min_bound, sycl::vec< f32, 2 > max_bound)
template<>
sycl::vec< f32, 3 >	mock_value (std::mt19937 &eng, sycl::vec< f32, 3 > min_bound, sycl::vec< f32, 3 > max_bound)
template<>
sycl::vec< f32, 4 >	mock_value (std::mt19937 &eng, sycl::vec< f32, 4 > min_bound, sycl::vec< f32, 4 > max_bound)
template<>
sycl::vec< f32, 8 >	mock_value (std::mt19937 &eng, sycl::vec< f32, 8 > min_bound, sycl::vec< f32, 8 > max_bound)
template<>
sycl::vec< f32, 16 >	mock_value (std::mt19937 &eng, sycl::vec< f32, 16 > min_bound, sycl::vec< f32, 16 > max_bound)
template<>
u8	mock_value (std::mt19937_64 &eng, u8 min_bound, u8 max_bound)
template<>
u16	mock_value (std::mt19937_64 &eng, u16 min_bound, u16 max_bound)
template<>
u32	mock_value (std::mt19937_64 &eng, u32 min_bound, u32 max_bound)
template<>
u64	mock_value (std::mt19937_64 &eng, u64 min_bound, u64 max_bound)
template<>
i8	mock_value (std::mt19937_64 &eng, i8 min_bound, i8 max_bound)
template<>
i16	mock_value (std::mt19937_64 &eng, i16 min_bound, i16 max_bound)
template<>
i32	mock_value (std::mt19937_64 &eng, i32 min_bound, i32 max_bound)
template<>
i64	mock_value (std::mt19937_64 &eng, i64 min_bound, i64 max_bound)
template<>
f32	mock_value (std::mt19937_64 &eng, f32 min_bound, f32 max_bound)
template<>
f64	mock_value (std::mt19937_64 &eng, f64 min_bound, f64 max_bound)
template<>
sycl::vec< u64, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< u64, 2 > min_bound, sycl::vec< u64, 2 > max_bound)
template<>
sycl::vec< u64, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< u64, 3 > min_bound, sycl::vec< u64, 3 > max_bound)
template<>
sycl::vec< u64, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< u64, 4 > min_bound, sycl::vec< u64, 4 > max_bound)
template<>
sycl::vec< u64, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< u64, 8 > min_bound, sycl::vec< u64, 8 > max_bound)
template<>
sycl::vec< u64, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< u64, 16 > min_bound, sycl::vec< u64, 16 > max_bound)
template<>
sycl::vec< u32, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< u32, 2 > min_bound, sycl::vec< u32, 2 > max_bound)
template<>
sycl::vec< u32, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< u32, 3 > min_bound, sycl::vec< u32, 3 > max_bound)
template<>
sycl::vec< u32, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< u32, 4 > min_bound, sycl::vec< u32, 4 > max_bound)
template<>
sycl::vec< u32, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< u32, 8 > min_bound, sycl::vec< u32, 8 > max_bound)
template<>
sycl::vec< u32, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< u32, 16 > min_bound, sycl::vec< u32, 16 > max_bound)
template<>
sycl::vec< u16, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< u16, 2 > min_bound, sycl::vec< u16, 2 > max_bound)
template<>
sycl::vec< u16, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< u16, 3 > min_bound, sycl::vec< u16, 3 > max_bound)
template<>
sycl::vec< u16, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< u16, 4 > min_bound, sycl::vec< u16, 4 > max_bound)
template<>
sycl::vec< u16, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< u16, 8 > min_bound, sycl::vec< u16, 8 > max_bound)
template<>
sycl::vec< u16, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< u16, 16 > min_bound, sycl::vec< u16, 16 > max_bound)
template<>
sycl::vec< u8, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< u8, 2 > min_bound, sycl::vec< u8, 2 > max_bound)
template<>
sycl::vec< u8, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< u8, 3 > min_bound, sycl::vec< u8, 3 > max_bound)
template<>
sycl::vec< u8, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< u8, 4 > min_bound, sycl::vec< u8, 4 > max_bound)
template<>
sycl::vec< u8, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< u8, 8 > min_bound, sycl::vec< u8, 8 > max_bound)
template<>
sycl::vec< u8, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< u8, 16 > min_bound, sycl::vec< u8, 16 > max_bound)
template<>
sycl::vec< i64, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< i64, 2 > min_bound, sycl::vec< i64, 2 > max_bound)
template<>
sycl::vec< i64, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< i64, 3 > min_bound, sycl::vec< i64, 3 > max_bound)
template<>
sycl::vec< i64, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< i64, 4 > min_bound, sycl::vec< i64, 4 > max_bound)
template<>
sycl::vec< i64, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< i64, 8 > min_bound, sycl::vec< i64, 8 > max_bound)
template<>
sycl::vec< i64, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< i64, 16 > min_bound, sycl::vec< i64, 16 > max_bound)
template<>
sycl::vec< i32, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< i32, 2 > min_bound, sycl::vec< i32, 2 > max_bound)
template<>
sycl::vec< i32, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< i32, 3 > min_bound, sycl::vec< i32, 3 > max_bound)
template<>
sycl::vec< i32, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< i32, 4 > min_bound, sycl::vec< i32, 4 > max_bound)
template<>
sycl::vec< i32, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< i32, 8 > min_bound, sycl::vec< i32, 8 > max_bound)
template<>
sycl::vec< i32, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< i32, 16 > min_bound, sycl::vec< i32, 16 > max_bound)
template<>
sycl::vec< i16, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< i16, 2 > min_bound, sycl::vec< i16, 2 > max_bound)
template<>
sycl::vec< i16, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< i16, 3 > min_bound, sycl::vec< i16, 3 > max_bound)
template<>
sycl::vec< i16, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< i16, 4 > min_bound, sycl::vec< i16, 4 > max_bound)
template<>
sycl::vec< i16, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< i16, 8 > min_bound, sycl::vec< i16, 8 > max_bound)
template<>
sycl::vec< i16, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< i16, 16 > min_bound, sycl::vec< i16, 16 > max_bound)
template<>
sycl::vec< i8, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< i8, 2 > min_bound, sycl::vec< i8, 2 > max_bound)
template<>
sycl::vec< i8, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< i8, 3 > min_bound, sycl::vec< i8, 3 > max_bound)
template<>
sycl::vec< i8, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< i8, 4 > min_bound, sycl::vec< i8, 4 > max_bound)
template<>
sycl::vec< i8, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< i8, 8 > min_bound, sycl::vec< i8, 8 > max_bound)
template<>
sycl::vec< i8, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< i8, 16 > min_bound, sycl::vec< i8, 16 > max_bound)
template<>
sycl::vec< f64, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< f64, 2 > min_bound, sycl::vec< f64, 2 > max_bound)
template<>
sycl::vec< f64, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< f64, 3 > min_bound, sycl::vec< f64, 3 > max_bound)
template<>
sycl::vec< f64, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< f64, 4 > min_bound, sycl::vec< f64, 4 > max_bound)
template<>
sycl::vec< f64, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< f64, 8 > min_bound, sycl::vec< f64, 8 > max_bound)
template<>
sycl::vec< f64, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< f64, 16 > min_bound, sycl::vec< f64, 16 > max_bound)
template<>
sycl::vec< f32, 2 >	mock_value (std::mt19937_64 &eng, sycl::vec< f32, 2 > min_bound, sycl::vec< f32, 2 > max_bound)
template<>
sycl::vec< f32, 3 >	mock_value (std::mt19937_64 &eng, sycl::vec< f32, 3 > min_bound, sycl::vec< f32, 3 > max_bound)
template<>
sycl::vec< f32, 4 >	mock_value (std::mt19937_64 &eng, sycl::vec< f32, 4 > min_bound, sycl::vec< f32, 4 > max_bound)
template<>
sycl::vec< f32, 8 >	mock_value (std::mt19937_64 &eng, sycl::vec< f32, 8 > min_bound, sycl::vec< f32, 8 > max_bound)
template<>
sycl::vec< f32, 16 >	mock_value (std::mt19937_64 &eng, sycl::vec< f32, 16 > min_bound, sycl::vec< f32, 16 > max_bound)
REDUCTION_IMPL	get_default_reduction_impl ()
REDUCTION_IMPL	reduction_impl_from_params (const std::string &impl)
shamalgs::impl_param	reduction_impl_to_params (const REDUCTION_IMPL &impl)
EXSCAN_IN_PLACE_IMPL	get_default_scan_exclusive_sum_in_place_impl ()
EXSCAN_IN_PLACE_IMPL	scan_exclusive_sum_in_place_impl_from_params (const std::string &impl)
shamalgs::impl_param	scan_exclusive_sum_in_place_impl_to_params (const EXSCAN_IN_PLACE_IMPL &impl)
template void	scan_exclusive_sum_in_place< u32 > (sham::DeviceBuffer< u32 > &buf1, u32 len)
template<class T , class Comp >
void	internal_segmented_sort_in_place (sham::DeviceBuffer< T > &buf, const sham::DeviceBuffer< u32 > &offsets, Comp &&comp)
template<>
void	segmented_sort_in_place< u32_2 > (sham::DeviceBuffer< u32_2 > &buf, const sham::DeviceBuffer< u32 > &offsets)
template<>
void	segmented_sort_in_place< u32 > (sham::DeviceBuffer< u32 > &buf, const sham::DeviceBuffer< u32 > &offsets)
template void	sort_by_key_pow2_len (sycl::queue &q, sycl::buffer< u32 > &buf_key, sycl::buffer< u32 > &buf_values, u32 len)
template void	sort_by_key_pow2_len (sycl::queue &q, sycl::buffer< u64 > &buf_key, sycl::buffer< u32 > &buf_values, u32 len)
template void	sort_by_key_pow2_len (const sham::DeviceScheduler_ptr &sched, sham::DeviceBuffer< u32 > &buf_key, sham::DeviceBuffer< u32 > &buf_values, u32 len)
template void	sort_by_key_pow2_len (const sham::DeviceScheduler_ptr &sched, sham::DeviceBuffer< u64 > &buf_key, sham::DeviceBuffer< u32 > &buf_values, u32 len)
template void	sort_by_key_pow2_len (const sham::DeviceScheduler_ptr &sched, sham::DeviceBuffer< f64 > &buf_key, sham::DeviceBuffer< f64 > &buf_values, u32 len)
template void	sort_by_key_pow2_len (const sham::DeviceScheduler_ptr &sched, sham::DeviceBuffer< f32 > &buf_key, sham::DeviceBuffer< f32 > &buf_values, u32 len)

Variables
IS_ALL_TRUE_IMPL	is_all_true_impl = IS_ALL_TRUE_IMPL::HOST
REDUCTION_IMPL	reduction_impl = get_default_reduction_impl()
EXSCAN_IN_PLACE_IMPL	scan_exclusive_sum_in_place_impl = get_default_scan_exclusive_sum_in_place_impl()

Detailed Description

namespace for primitive algorithm (e.g. sort, scan, reductions, ...)

Enumeration Type Documentation

EXSCAN_IN_PLACE_IMPL

enum class shamalgs::primitives::EXSCAN_IN_PLACE_IMPL : u32

strong

Definition at line 107 of file scan_exclusive_sum_in_place.cpp.

IS_ALL_TRUE_IMPL

enum class shamalgs::primitives::IS_ALL_TRUE_IMPL : u32

strong

Definition at line 69 of file is_all_true.cpp.

REDUCTION_IMPL

enum class shamalgs::primitives::REDUCTION_IMPL : u32

strong

Definition at line 31 of file reduction.cpp.

Function Documentation

append_subset_to()

template<class T >

void shamalgs::primitives::append_subset_to	(	const sham::DeviceBuffer< T > &	buf,
		const sham::DeviceBuffer< u32 > &	idxs_buf,
		u32	nvar,
		sham::DeviceBuffer< T > &	buf_other,
		u32	start_enque
	)

Appends a subset of elements from one buffer to another.

The elements to append are specified by indices in idxs_buf. The source buffer buf is treated as an array of objects, each with nvar variables. The elements are appended to buf_other.

Template Parameters

T	The type of data in the buffers.

Parameters

buf	The source buffer.
idxs_buf	A buffer of indices specifying which objects to copy from buf.
nvar	The number of variables per object.
buf_other	The destination buffer to which the subset will be appended.
start_enque	The starting index in buf_other where the subset will be appended.

Definition at line 45 of file append_subset_to.cpp.

Here is the call graph for this function:

binary_range_search()

template<class Tkey >

constexpr void shamalgs::primitives::binary_range_search ( const Tkey *__restrict__  key, u32  first, u32  last, const Tkey &  value_min, const Tkey &  value_max, u32 &  inf, u32 &  sup  )

constexpr

Find the range of indices for which key[inf] <= value_min <= value_max <= key[sup].

Precondition

key is sorted in ascending order

value_min <= value_max

first < last

key[first] <= value_min

key[last - 1] >= value_max

Template Parameters

Tkey	Type of the key

Parameters

key	Array of keys
first	First index of the range
last	Last index of the range
value_min	Lower bound of the range
value_max	Upper bound of the range
inf	Index of the first element of the range
sup	Index of the last element of the range

// Example usage with a sorted array
std::vector<int> keys = {1, 3, 5, 7, 9, 11, 13, 15};
u32 inf, sup;
 
// Find range of elements between 5 and 11 (inclusive)
binary_range_search(keys.data(), 0, keys.size(), 5, 11, inf, sup);
// inf = 2, sup = 5 (elements 5, 7, 9, 11 at indices 2, 3, 4, 5)
 
// Find range with duplicates
std::vector<int> keys2 = {1, 2, 2, 2, 3, 4, 4, 5};
binary_range_search(keys2.data(), 0, keys2.size(), 2, 4, inf, sup);
// inf = 1, sup = 6 (elements 2, 2, 2, 3, 4, 4 at indices 1-6)

Definition at line 59 of file binary_range_search.hpp.

Here is the call graph for this function:

binary_search_lower_bound()

template<class Tkey >

constexpr u32 shamalgs::primitives::binary_search_lower_bound ( const Tkey *__restrict__  key, u32  first, u32  last, const Tkey &  value  )

constexpr

GPU compatible implementation of std::lower_bound.

Definition at line 25 of file lower_bound.hpp.

binary_search_upper_bound()

template<class Tkey >

constexpr u32 shamalgs::primitives::binary_search_upper_bound ( const Tkey *__restrict__  key, u32  first, u32  last, const Tkey &  value  )

constexpr

GPU compatible implementation of std::upper_bound.

Definition at line 25 of file upper_bound.hpp.

compute_histogram()

template<class T , class Tbins , class... Targs, class Tfunctor >

sham::DeviceBuffer< T > shamalgs::primitives::compute_histogram ( const sham::DeviceScheduler_ptr &  dev_sched, const sham::DeviceBuffer< Tbins > &  bin_edge_inf, const sham::DeviceBuffer< Tbins > &  bin_edge_sup, size_t  element_count, Tfunctor &&  functor, const sham::DeviceBuffer< Targs > &...  input_data  )

inline

Definition at line 359 of file compute_histogram.hpp.

compute_histogram_basic()

template<class T >

sham::DeviceBuffer< T > shamalgs::primitives::compute_histogram_basic ( const sham::DeviceScheduler_ptr &  dev_sched, const sham::DeviceBuffer< T > &  bin_edge_inf, const sham::DeviceBuffer< T > &  bin_edge_sup, const sham::DeviceBuffer< T > &  positions  )

inline

Definition at line 430 of file compute_histogram.hpp.

dot_sum() [1/2]

template<class T >

shambase::VecComponent< T > shamalgs::primitives::dot_sum ( sham::DeviceBuffer< T > &  buf1 )

inline

Definition at line 45 of file dot_sum.hpp.

dot_sum() [2/2]

template<class T >

shambase::VecComponent< T > shamalgs::primitives::dot_sum	(	sham::DeviceBuffer< T > &	buf1,
		u32	start_id,
		u32	end_id
	)

Compute the sum of dot products of elements in a device buffer with themselves.

Template Parameters

T	The data type of elements in the buffer (e.g., float, double, int).

Parameters

buf1	The input buffer containing the elements to sum.
start_id	The starting index (inclusive) of the range to sum.
end_id	The ending index (exclusive) of the range to sum.

Returns

shambase::VecComponent<T> The computed sum of dot products.

Example:

auto sched = shamsys::get_compute_Scheduler_ptr();
sham::DeviceBuffer<f64> values = ...;
f64 result = shamalgs::primitives::dot_sum(sched, values, 0, values.get_size());

Definition at line 26 of file dot_sum.cpp.

Here is the call graph for this function:

equals() [1/4]

template<class T >

bool shamalgs::primitives::equals ( const sham::DeviceScheduler_ptr &  dev_sched, sham::DeviceBuffer< T > &  buf1, sham::DeviceBuffer< T > &  buf2, u32  cnt  )

inline

Compare elements between two sham::DeviceBuffers for equality.

Performs element-wise comparison between two device buffers to determine if all corresponding elements are equal. The function compares the first cnt elements of each buffer and returns true only if all pairs match. This is the preferred method for USM-based buffer comparisons.

Template Parameters

T	Element type - must support equality comparison

Parameters

dev_sched	Device scheduler pointer for execution context
buf1	First device buffer to compare
buf2	Second device buffer to compare
cnt	Number of elements to compare from the beginning of each buffer

Returns

true if all compared elements are equal, false otherwise

Exceptions

std::invalid_argument

if either buffer is smaller than cnt

Note

Returns true immediately if both buffers are the same object

Uses parallel device execution for the comparison operation

// Example: Compare device buffers element-wise
auto sched = shamsys::get_compute_Scheduler_ptr();
std::vector<f32> data1 = {1.0f, 2.0f, 3.0f, 4.0f};
std::vector<f32> data2 = {1.0f, 2.0f, 3.0f, 4.0f};
 
sham::DeviceBuffer<f32> buffer1(data1.size(), sched);
sham::DeviceBuffer<f32> buffer2(data2.size(), sched);
 
// Initialize buffers with data
buffer1.copy_from_stdvec(data1);
buffer2.copy_from_stdvec(data2);
 
// Compare first 4 elements
bool are_equal = equals(sched, buffer1, buffer2, 4);
// Result: true (all elements match)
 
// Example with tolerance-based comparison
std::vector<f32> data3 = {1.0f, 2.0f, 3.1f, 4.0f};
sham::DeviceBuffer<f32> buffer3(data3.size(), sched);
buffer3.copy_from_stdvec(data3);
bool different = equals(sched, buffer1, buffer3, 4);
// Result: false (third element differs)

Definition at line 154 of file equals.hpp.

Here is the call graph for this function:

equals() [2/4]

template<class T >

bool shamalgs::primitives::equals ( const sham::DeviceScheduler_ptr &  q, sham::DeviceBuffer< T > &  buf1, sham::DeviceBuffer< T > &  buf2  )

inline

Compare all elements between two sham::DeviceBuffers for equality.

Performs element-wise comparison between two device buffers to determine if all corresponding elements are equal. This function automatically compares all elements if both buffers have the same size, or returns false if sizes differ.

Template Parameters

T	Element type - must support equality comparison

Parameters

q	Device scheduler pointer for execution context
buf1	First device buffer to compare
buf2	Second device buffer to compare

Returns

true if buffers have same size and all elements are equal, false otherwise

Note

Returns false immediately if buffer sizes differ

Convenience function that calls equals(q, buf1, buf2, buf1.get_size())

// Example: Compare entire device buffers
auto sched = shamsys::get_compute_Scheduler_ptr();
std::vector<i32> keys = {10, 20, 30, 40};
std::vector<i32> values = {10, 20, 30, 40};
 
sham::DeviceBuffer<i32> buffer1(keys.size(), sched);
sham::DeviceBuffer<i32> buffer2(values.size(), sched);
 
buffer1.copy_from_stdvec(keys);
buffer2.copy_from_stdvec(values);
 
// Compare entire buffers
bool are_equal = equals(sched, buffer1, buffer2);
// Result: true (all elements match)
 
// Example with different sizes
std::vector<i32> shorter = {10, 20};
sham::DeviceBuffer<i32> buffer3(shorter.size(), sched);
buffer3.copy_from_stdvec(shorter);
bool size_diff = equals(sched, buffer1, buffer3);
// Result: false (different sizes)

Definition at line 235 of file equals.hpp.

Here is the call graph for this function:

equals() [3/4]

template<class T >

bool shamalgs::primitives::equals	(	sycl::queue &	q,
		sycl::buffer< T > &	buf1,
		sycl::buffer< T > &	buf2
	)

Compare all elements between two sycl::buffers for equality.

Performs element-wise comparison between two buffers to determine if all corresponding elements are equal. This function automatically compares all elements if both buffers have the same size, or returns false if sizes differ.

Template Parameters

T	Element type - must support equality comparison

Parameters

q	sycl::queue for device execution
buf1	First buffer to compare
buf2	Second buffer to compare

Returns

true if buffers have same size and all elements are equal, false otherwise

Note

Returns false immediately if buffer sizes differ

Convenience function that calls equals(q, buf1, buf2, buf1.size())

Deprecated:

Use equals with sham::DeviceBuffer instead

// Example: Compare entire sycl buffers
std::vector<f64> data1 = {1.0, 2.0, 3.0};
std::vector<f64> data2 = {1.0, 2.0, 3.0};
sycl::buffer<f64> buffer1(data1);
sycl::buffer<f64> buffer2(data2);
sycl::queue q;
 
// Compare entire buffers
bool are_equal = equals(q, buffer1, buffer2);
// Result: true (all elements match)

Definition at line 280 of file equals.hpp.

Here is the call graph for this function:

equals() [4/4]

template<class T >

bool shamalgs::primitives::equals	(	sycl::queue &	q,
		sycl::buffer< T > &	buf1,
		sycl::buffer< T > &	buf2,
		u32	cnt
	)

Compare elements between two sycl::buffers for equality.

Performs element-wise comparison between two buffers to determine if all corresponding elements are equal. The function compares the first cnt elements of each buffer and returns true only if all pairs match.

Template Parameters

T	Element type - must support equality comparison

Parameters

q	sycl::queue for device execution
buf1	First buffer to compare
buf2	Second buffer to compare
cnt	Number of elements to compare from the beginning of each buffer

Returns

true if all compared elements are equal, false otherwise

Exceptions

std::invalid_argument

if either buffer is smaller than cnt

Deprecated:

Use equals(const sham::DeviceScheduler_ptr &, sham::DeviceBuffer<T> &, sham::DeviceBuffer<T> &, u32 ) instead.

// Example: Compare arrays element-wise
std::vector<i32> data1 = {1, 2, 3, 4, 5};
std::vector<i32> data2 = {1, 2, 3, 4, 5};
sycl::buffer<i32> buffer1(data1);
sycl::buffer<i32> buffer2(data2);
sycl::queue q;
 
// Compare first 5 elements
bool are_equal = equals(q, buffer1, buffer2, 5);
// Result: true (all elements match)
 
// Example with different data
std::vector<i32> data3 = {1, 2, 0, 4, 5};
sycl::buffer<i32> buffer3(data3);
bool different = equals(q, buffer1, buffer3, 5);
// Result: false (third element differs)

Definition at line 77 of file equals.hpp.

Here is the call graph for this function:

equals_ptr()

template<class T >

bool shamalgs::primitives::equals_ptr	(	sycl::queue &	q,
		const std::unique_ptr< sycl::buffer< T > > &	buf1,
		const std::unique_ptr< sycl::buffer< T > > &	buf2
	)

Compare all elements between two unique_ptr-wrapped sycl::buffers.

Performs element-wise comparison between two unique_ptr-wrapped buffers to determine if all corresponding elements are equal. The function handles null pointer cases and compares all elements if both pointers are valid and buffer sizes match.

Template Parameters

T	Element type - must support equality comparison

Parameters

q	sycl::queue for device execution
buf1	First unique_ptr-wrapped buffer to compare
buf2	Second unique_ptr-wrapped buffer to compare

Returns

true if both are null or all elements are equal, false otherwise

Note

Returns false if only one pointer is null

Returns true if both pointers are null

Uses equals(q, *buf1, *buf2) for actual comparison

Deprecated:

Use sham::DeviceBuffer-based functions instead

// Example: Compare optional buffers entirely
sycl::queue q;
std::vector<f32> data1 = {1.0f, 2.0f, 3.0f};
std::vector<f32> data2 = {1.0f, 2.0f, 3.0f};
 
auto buf1 = std::make_unique<sycl::buffer<f32>>(data1);
auto buf2 = std::make_unique<sycl::buffer<f32>>(data2);
 
// Compare entire buffers
bool are_equal = equals_ptr(q, buf1, buf2);
// Result: true (all elements match)
 
// Example with one null pointer
std::unique_ptr<sycl::buffer<f32>> null_buf;
bool one_null = equals_ptr(q, buf1, null_buf);
// Result: false (only one is null)

Definition at line 386 of file equals.hpp.

Here is the call graph for this function:

equals_ptr_s()

template<class T >

bool shamalgs::primitives::equals_ptr_s	(	sycl::queue &	q,
		const std::unique_ptr< sycl::buffer< T > > &	buf1,
		const std::unique_ptr< sycl::buffer< T > > &	buf2,
		u32	cnt
	)

Compare elements between two unique_ptr-wrapped sycl::buffers with count.

Performs element-wise comparison between two unique_ptr-wrapped buffers to determine if all corresponding elements are equal. The function handles null pointer cases and compares the first cnt elements if both pointers are valid.

Template Parameters

T	Element type - must support equality comparison

Parameters

q	sycl::queue for device execution
buf1	First unique_ptr-wrapped buffer to compare
buf2	Second unique_ptr-wrapped buffer to compare
cnt	Number of elements to compare from the beginning of each buffer

Returns

true if both are null or all compared elements are equal, false otherwise

Note

Returns false if only one pointer is null

Returns true if both pointers are null

Uses equals(q, *buf1, *buf2, cnt) for actual comparison

Deprecated:

Use sham::DeviceBuffer-based functions instead

// Example: Compare optional buffers with count
sycl::queue q;
std::vector<i32> data1 = {1, 2, 3, 4, 5};
std::vector<i32> data2 = {1, 2, 3, 4, 5};
 
auto buf1 = std::make_unique<sycl::buffer<i32>>(data1);
auto buf2 = std::make_unique<sycl::buffer<i32>>(data2);
 
// Compare first 3 elements
bool are_equal = equals_ptr_s(q, buf1, buf2, 3);
// Result: true (first 3 elements match)
 
// Example with null pointers
std::unique_ptr<sycl::buffer<i32>> null_buf1, null_buf2;
bool both_null = equals_ptr_s(q, null_buf1, null_buf2, 0);
// Result: true (both are null)

Definition at line 329 of file equals.hpp.

Here is the call graph for this function:

fill_buffer_index()

void shamalgs::primitives::fill_buffer_index	(	sham::DeviceBuffer< u32 > &	buf,
		u32	len
	)

Fills a buffer with sequential indices, such that buf[i] = i.

The buffer must have a size of at least len. If len is 0, the function has no effect.

Parameters

buf	The buffer to fill.
len	The number of elements to fill from the start of the buffer.

Exceptions

std::invalid_argument

if buf.get_size() < len.

Definition at line 23 of file gen_buffer_index.cpp.

Here is the call graph for this function:

flatten_buffer()

template<class Tvec , sham::USMKindTarget target>

sham::DeviceBuffer< typename shambase::VectorProperties< Tvec >::component_type, target > shamalgs::primitives::flatten_buffer ( const sham::DeviceBuffer< Tvec, target > &  buffer )

inline

Flatten a buffer of vector type into a buffer of scalar type.

The buffer is flattened in row-major order, i.e. the first element of the first vector is stored first, then the second element of the first vector, and so on.

Parameters

buffer

The buffer to flatten.

Returns

The flattened buffer.

Definition at line 40 of file flatten.hpp.

Here is the call graph for this function:

gen_buffer_index()

sham::DeviceBuffer< u32 > shamalgs::primitives::gen_buffer_index	(	sham::DeviceScheduler_ptr	sched,
		u32	len
	)

Generates a buffer where buf[i] = i.

Parameters

sched	The scheduler to run on.
len	Length of the buffer to generate.

Returns

sham::DeviceBuffer<u32> The generated buffer containing sequential indices.

Definition at line 39 of file gen_buffer_index.cpp.

Here is the call graph for this function:

get_default_reduction_impl()

REDUCTION_IMPL shamalgs::primitives::get_default_reduction_impl

(

)

Definition at line 40 of file reduction.cpp.

get_default_scan_exclusive_sum_in_place_impl()

EXSCAN_IN_PLACE_IMPL shamalgs::primitives::get_default_scan_exclusive_sum_in_place_impl

(

)

Definition at line 120 of file scan_exclusive_sum_in_place.cpp.

internal_segmented_sort_in_place()

template<class T , class Comp >

void shamalgs::primitives::internal_segmented_sort_in_place	(	sham::DeviceBuffer< T > &	buf,
		const sham::DeviceBuffer< u32 > &	offsets,
		Comp &&	comp
	)

Definition at line 168 of file segmented_sort_in_place.cpp.

is_all_true() [1/2]

template<class T >

bool shamalgs::primitives::is_all_true	(	sham::DeviceBuffer< T > &	buf,
		u32	cnt
	)

Check if all elements in a sham::DeviceBuffer are non-zero.

Performs a boolean reduction operation to determine if every element in the device buffer is non-zero (logically true). The function returns true only if all elements satisfy the condition element != 0.

Template Parameters

T	Element type - must support comparison with zero

Parameters

buf	Device buffer containing the elements to check
cnt	Number of elements to check from the beginning of the buffer

Returns

true if all elements are non-zero, false otherwise

Note

Currently implemented by copying to sycl::buffer and using CPU processing

The function only checks the first cnt elements of the buffer

// Example: Check device buffer elements
std::vector<u8> flags = {1, 1, 1, 1};
sham::DeviceBuffer<u8> device_buffer(flags);
 
// Check all flags
bool all_set = is_all_true(device_buffer, flags.size());
// Result: true (all flags are set)
 
// Example with boolean conditions
std::vector<i32> conditions = {5 > 0, 3 > 0, 0 > 0, 1 > 0};
sham::DeviceBuffer<i32> cond_buffer(conditions);
bool all_conditions_met = is_all_true(cond_buffer, conditions.size());
// Result: false (third condition evaluates to 0)

Definition at line 109 of file is_all_true.cpp.

Here is the call graph for this function:

is_all_true() [2/2]

template<class T >

bool shamalgs::primitives::is_all_true	(	sycl::buffer< T > &	buf,
		u32	cnt
	)

Check if all elements in a sycl::buffer are non-zero.

Performs a boolean reduction operation to determine if every element in the buffer is non-zero (logically true). The function returns true only if all elements satisfy the condition element != 0.

Template Parameters

T	Element type - must support comparison with zero

Parameters

buf	Buffer containing the elements to check
cnt	Number of elements to check from the beginning of the buffer

Returns

true if all elements are non-zero, false otherwise

Note

Currently implemented on CPU but marked for GPU optimization

The function only checks the first cnt elements of the buffer

Deprecated:

Use is_all_true(sham::DeviceBuffer<T>&, u32) instead.

// Example: Check if all elements are non-zero
std::vector<i32> data = {1, 2, 3, 4, 5};
sycl::buffer<i32> buffer(data);
 
// Check all elements
bool all_nonzero = is_all_true(buffer, data.size());
// Result: true (all elements are non-zero)
 
// Example with zeros
std::vector<i32> data_with_zero = {1, 0, 3, 4};
sycl::buffer<i32> buffer2(data_with_zero);
bool has_zero = is_all_true(buffer2, data_with_zero.size());
// Result: false (contains zero)

Definition at line 124 of file is_all_true.cpp.

is_all_true_impl_from_params()

IS_ALL_TRUE_IMPL shamalgs::primitives::is_all_true_impl_from_params ( const std::string &  impl )

inline

Definition at line 72 of file is_all_true.cpp.

is_all_true_impl_to_params()

shamalgs::impl_param shamalgs::primitives::is_all_true_impl_to_params ( const IS_ALL_TRUE_IMPL &  impl )

inline

Definition at line 84 of file is_all_true.cpp.

linspace()

template<typename Tval >

sham::DeviceBuffer< Tval > shamalgs::primitives::linspace	(	Tval	Rmin,
		Tval	Rmax,
		u32	N
	)

Create an array of N values between two values.

Template Parameters

Tval	value type in the sequence (same type as bounds)

Parameters

Rmin	lower bound of the sequence
Rmax	upper bound of the sequence
N	number of samples to generate

Returns

a DeviceBuffer containing the sequence

TODO: move to GPU

Definition at line 37 of file linspace.hpp.

max()

template<class T >

T shamalgs::primitives::max	(	const sham::DeviceScheduler_ptr &	sched,
		const sham::DeviceBuffer< T > &	buf1,
		u32	start_id,
		u32	end_id
	)

Find the maximum element in a device buffer within a specified range.

This function finds the maximum value among all elements in the buffer between start_id (inclusive) and end_id (exclusive). The computation is performed on the device using the provided scheduler.

Template Parameters

T	The data type of elements in the buffer (e.g., float, double, int).

Parameters

sched	The device scheduler to run on.
buf1	The input buffer containing the elements to search.
start_id	The starting index (inclusive) of the range to search.
end_id	The ending index (exclusive) of the range to search.

Returns

T The maximum value found in the specified range.

Precondition

start_id < end_id (range must be non-empty)

end_id <= buf1.get_size()

Example:

auto sched = shamsys::get_compute_Scheduler_ptr();
 
sham::DeviceBuffer<double> values = ...;
u32 start = 0;
u32 end = values.get_size();
 
double maximum = shamalgs::primitives::max(sched, values, start, end);

values = {5.0, 2.0, 8.0, 1.0, 6.0}, start = 1, end = 4 result = 8.0 (maximum of {2.0, 8.0, 1.0})

Definition at line 155 of file reduction.cpp.

Here is the call graph for this function:

min()

template<class T >

T shamalgs::primitives::min	(	const sham::DeviceScheduler_ptr &	sched,
		const sham::DeviceBuffer< T > &	buf1,
		u32	start_id,
		u32	end_id
	)

Find the minimum element in a device buffer within a specified range.

This function finds the minimum value among all elements in the buffer between start_id (inclusive) and end_id (exclusive). The computation is performed on the device using the provided scheduler.

Template Parameters

T	The data type of elements in the buffer (e.g., float, double, int).

Parameters

sched	The device scheduler to run on.
buf1	The input buffer containing the elements to search.
start_id	The starting index (inclusive) of the range to search.
end_id	The ending index (exclusive) of the range to search.

Returns

T The minimum value found in the specified range.

Precondition

start_id < end_id (range must be non-empty)

end_id <= buf1.get_size()

Example:

auto sched = shamsys::get_compute_Scheduler_ptr();
 
sham::DeviceBuffer<double> values = ...;
u32 start = 0;
u32 end = values.get_size();
 
double minimum = shamalgs::primitives::min(sched, values, start, end);

values = {5.0, 2.0, 8.0, 1.0, 6.0}, start = 1, end = 4 result = 1.0 (minimum of {2.0, 8.0, 1.0})

Definition at line 130 of file reduction.cpp.

Here is the call graph for this function:

mock_value() [1/122]

template<class T , class Engine = std::mt19937>

T shamalgs::primitives::mock_value ( Engine &  eng )

inline

Generates a random mock value using default bounds.

This function generates a random value of type T using the default minimum and maximum bounds defined by shambase::VectorProperties<T>. This is a convenience function that automatically determines appropriate bounds based on the type T.

Template Parameters

T	The type of value to generate (supports primitive types and SYCL vectors)

Parameters

eng	The random number generator engine to use

Returns

A random value of type T within the default bounds for the type

#include <random>
#include "shamalgs/primitives/mock_value.hpp"
 
// Initialize random number generator
std::mt19937 rng(42);
 
// Generate random values with default bounds
i32 random_int = shamalgs::mock_value<i32>(rng);
f64 random_double = shamalgs::mock_value<f64>(rng);
sycl::vec<f32, 2> random_vec2 = shamalgs::mock_value<sycl::vec<f32, 2>>(rng);

Definition at line 90 of file mock_value.hpp.

mock_value() [2/122]

template<class T , class Engine = std::mt19937>

T shamalgs::primitives::mock_value	(	Engine &	eng,
		T	min_bound,
		T	max_bound
	)

Generates a random mock value within specified bounds.

This function generates a random value of type T within the range [min_bound, max_bound]. For integer types, the distribution is uniform over the inclusive range. For floating-point types, the distribution is uniform over the range [min_bound, max_bound). For SYCL vectors, each component is generated independently using the same bounds.

Template Parameters

T	The type of value to generate (supports primitive types and SYCL vectors)

Parameters

eng	The random number generator engine to use
min_bound	The minimum bound for the generated value
max_bound	The maximum bound for the generated value

Returns

A random value of type T within the specified bounds

#include <random>
#include "shamalgs/primitives/mock_value.hpp"
 
// Initialize random number generator
std::mt19937 rng(42);
 
// Generate random integers
i32 random_int = shamalgs::mock_value(rng, i32{0}, i32{100});
 
// Generate random floats
f32 random_float = shamalgs::mock_value(rng, f32{0.0f}, f32{1.0f});
 
// Generate random SYCL vectors
sycl::vec<f32, 3> random_vec = shamalgs::mock_value(
    rng,
    sycl::vec<f32, 3>{0.0f, 0.0f, 0.0f},
    sycl::vec<f32, 3>{1.0f, 1.0f, 1.0f}
);

mock_value() [3/122]

template<>

f32 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		f32	min_bound,
		f32	max_bound
	)

Definition at line 70 of file mock_value.cpp.

mock_value() [4/122]

template<>

f64 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		f64	min_bound,
		f64	max_bound
	)

Definition at line 76 of file mock_value.cpp.

mock_value() [5/122]

template<>

i16 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		i16	min_bound,
		i16	max_bound
	)

Definition at line 52 of file mock_value.cpp.

mock_value() [6/122]

template<>

i32 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		i32	min_bound,
		i32	max_bound
	)

Definition at line 58 of file mock_value.cpp.

mock_value() [7/122]

template<>

i64 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		i64	min_bound,
		i64	max_bound
	)

Definition at line 64 of file mock_value.cpp.

mock_value() [8/122]

template<>

i8 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		i8	min_bound,
		i8	max_bound
	)

Definition at line 46 of file mock_value.cpp.

mock_value() [9/122]

template<>

sycl::vec< f32, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f32, 16 >	min_bound,
		sycl::vec< f32, 16 >	max_bound
	)

Definition at line 162 of file mock_value.cpp.

mock_value() [10/122]

template<>

sycl::vec< f32, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f32, 2 >	min_bound,
		sycl::vec< f32, 2 >	max_bound
	)

Definition at line 162 of file mock_value.cpp.

mock_value() [11/122]

template<>

sycl::vec< f32, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f32, 3 >	min_bound,
		sycl::vec< f32, 3 >	max_bound
	)

Definition at line 162 of file mock_value.cpp.

mock_value() [12/122]

template<>

sycl::vec< f32, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f32, 4 >	min_bound,
		sycl::vec< f32, 4 >	max_bound
	)

Definition at line 162 of file mock_value.cpp.

mock_value() [13/122]

template<>

sycl::vec< f32, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f32, 8 >	min_bound,
		sycl::vec< f32, 8 >	max_bound
	)

Definition at line 162 of file mock_value.cpp.

mock_value() [14/122]

template<>

sycl::vec< f64, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f64, 16 >	min_bound,
		sycl::vec< f64, 16 >	max_bound
	)

Definition at line 161 of file mock_value.cpp.

mock_value() [15/122]

template<>

sycl::vec< f64, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f64, 2 >	min_bound,
		sycl::vec< f64, 2 >	max_bound
	)

Definition at line 161 of file mock_value.cpp.

mock_value() [16/122]

template<>

sycl::vec< f64, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f64, 3 >	min_bound,
		sycl::vec< f64, 3 >	max_bound
	)

Definition at line 161 of file mock_value.cpp.

mock_value() [17/122]

template<>

sycl::vec< f64, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f64, 4 >	min_bound,
		sycl::vec< f64, 4 >	max_bound
	)

Definition at line 161 of file mock_value.cpp.

mock_value() [18/122]

template<>

sycl::vec< f64, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< f64, 8 >	min_bound,
		sycl::vec< f64, 8 >	max_bound
	)

Definition at line 161 of file mock_value.cpp.

mock_value() [19/122]

template<>

sycl::vec< i16, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i16, 16 >	min_bound,
		sycl::vec< i16, 16 >	max_bound
	)

Definition at line 158 of file mock_value.cpp.

mock_value() [20/122]

template<>

sycl::vec< i16, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i16, 2 >	min_bound,
		sycl::vec< i16, 2 >	max_bound
	)

Definition at line 158 of file mock_value.cpp.

mock_value() [21/122]

template<>

sycl::vec< i16, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i16, 3 >	min_bound,
		sycl::vec< i16, 3 >	max_bound
	)

Definition at line 158 of file mock_value.cpp.

mock_value() [22/122]

template<>

sycl::vec< i16, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i16, 4 >	min_bound,
		sycl::vec< i16, 4 >	max_bound
	)

Definition at line 158 of file mock_value.cpp.

mock_value() [23/122]

template<>

sycl::vec< i16, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i16, 8 >	min_bound,
		sycl::vec< i16, 8 >	max_bound
	)

Definition at line 158 of file mock_value.cpp.

mock_value() [24/122]

template<>

sycl::vec< i32, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i32, 16 >	min_bound,
		sycl::vec< i32, 16 >	max_bound
	)

Definition at line 157 of file mock_value.cpp.

mock_value() [25/122]

template<>

sycl::vec< i32, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i32, 2 >	min_bound,
		sycl::vec< i32, 2 >	max_bound
	)

Definition at line 157 of file mock_value.cpp.

mock_value() [26/122]

template<>

sycl::vec< i32, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i32, 3 >	min_bound,
		sycl::vec< i32, 3 >	max_bound
	)

Definition at line 157 of file mock_value.cpp.

mock_value() [27/122]

template<>

sycl::vec< i32, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i32, 4 >	min_bound,
		sycl::vec< i32, 4 >	max_bound
	)

Definition at line 157 of file mock_value.cpp.

mock_value() [28/122]

template<>

sycl::vec< i32, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i32, 8 >	min_bound,
		sycl::vec< i32, 8 >	max_bound
	)

Definition at line 157 of file mock_value.cpp.

mock_value() [29/122]

template<>

sycl::vec< i64, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i64, 16 >	min_bound,
		sycl::vec< i64, 16 >	max_bound
	)

Definition at line 156 of file mock_value.cpp.

mock_value() [30/122]

template<>

sycl::vec< i64, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i64, 2 >	min_bound,
		sycl::vec< i64, 2 >	max_bound
	)

Definition at line 156 of file mock_value.cpp.

mock_value() [31/122]

template<>

sycl::vec< i64, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i64, 3 >	min_bound,
		sycl::vec< i64, 3 >	max_bound
	)

Definition at line 156 of file mock_value.cpp.

mock_value() [32/122]

template<>

sycl::vec< i64, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i64, 4 >	min_bound,
		sycl::vec< i64, 4 >	max_bound
	)

Definition at line 156 of file mock_value.cpp.

mock_value() [33/122]

template<>

sycl::vec< i64, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i64, 8 >	min_bound,
		sycl::vec< i64, 8 >	max_bound
	)

Definition at line 156 of file mock_value.cpp.

mock_value() [34/122]

template<>

sycl::vec< i8, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i8, 16 >	min_bound,
		sycl::vec< i8, 16 >	max_bound
	)

Definition at line 159 of file mock_value.cpp.

mock_value() [35/122]

template<>

sycl::vec< i8, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i8, 2 >	min_bound,
		sycl::vec< i8, 2 >	max_bound
	)

Definition at line 159 of file mock_value.cpp.

mock_value() [36/122]

template<>

sycl::vec< i8, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i8, 3 >	min_bound,
		sycl::vec< i8, 3 >	max_bound
	)

Definition at line 159 of file mock_value.cpp.

mock_value() [37/122]

template<>

sycl::vec< i8, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i8, 4 >	min_bound,
		sycl::vec< i8, 4 >	max_bound
	)

Definition at line 159 of file mock_value.cpp.

mock_value() [38/122]

template<>

sycl::vec< i8, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< i8, 8 >	min_bound,
		sycl::vec< i8, 8 >	max_bound
	)

Definition at line 159 of file mock_value.cpp.

mock_value() [39/122]

template<>

sycl::vec< u16, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u16, 16 >	min_bound,
		sycl::vec< u16, 16 >	max_bound
	)

Definition at line 153 of file mock_value.cpp.

mock_value() [40/122]

template<>

sycl::vec< u16, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u16, 2 >	min_bound,
		sycl::vec< u16, 2 >	max_bound
	)

Definition at line 153 of file mock_value.cpp.

mock_value() [41/122]

template<>

sycl::vec< u16, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u16, 3 >	min_bound,
		sycl::vec< u16, 3 >	max_bound
	)

Definition at line 153 of file mock_value.cpp.

mock_value() [42/122]

template<>

sycl::vec< u16, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u16, 4 >	min_bound,
		sycl::vec< u16, 4 >	max_bound
	)

Definition at line 153 of file mock_value.cpp.

mock_value() [43/122]

template<>

sycl::vec< u16, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u16, 8 >	min_bound,
		sycl::vec< u16, 8 >	max_bound
	)

Definition at line 153 of file mock_value.cpp.

mock_value() [44/122]

template<>

sycl::vec< u32, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u32, 16 >	min_bound,
		sycl::vec< u32, 16 >	max_bound
	)

Definition at line 152 of file mock_value.cpp.

mock_value() [45/122]

template<>

sycl::vec< u32, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u32, 2 >	min_bound,
		sycl::vec< u32, 2 >	max_bound
	)

Definition at line 152 of file mock_value.cpp.

mock_value() [46/122]

template<>

sycl::vec< u32, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u32, 3 >	min_bound,
		sycl::vec< u32, 3 >	max_bound
	)

Definition at line 152 of file mock_value.cpp.

mock_value() [47/122]

template<>

sycl::vec< u32, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u32, 4 >	min_bound,
		sycl::vec< u32, 4 >	max_bound
	)

Definition at line 152 of file mock_value.cpp.

mock_value() [48/122]

template<>

sycl::vec< u32, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u32, 8 >	min_bound,
		sycl::vec< u32, 8 >	max_bound
	)

Definition at line 152 of file mock_value.cpp.

mock_value() [49/122]

template<>

sycl::vec< u64, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u64, 16 >	min_bound,
		sycl::vec< u64, 16 >	max_bound
	)

Definition at line 151 of file mock_value.cpp.

mock_value() [50/122]

template<>

sycl::vec< u64, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u64, 2 >	min_bound,
		sycl::vec< u64, 2 >	max_bound
	)

Definition at line 151 of file mock_value.cpp.

mock_value() [51/122]

template<>

sycl::vec< u64, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u64, 3 >	min_bound,
		sycl::vec< u64, 3 >	max_bound
	)

Definition at line 151 of file mock_value.cpp.

mock_value() [52/122]

template<>

sycl::vec< u64, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u64, 4 >	min_bound,
		sycl::vec< u64, 4 >	max_bound
	)

Definition at line 151 of file mock_value.cpp.

mock_value() [53/122]

template<>

sycl::vec< u64, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u64, 8 >	min_bound,
		sycl::vec< u64, 8 >	max_bound
	)

Definition at line 151 of file mock_value.cpp.

mock_value() [54/122]

template<>

sycl::vec< u8, 16 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u8, 16 >	min_bound,
		sycl::vec< u8, 16 >	max_bound
	)

Definition at line 154 of file mock_value.cpp.

mock_value() [55/122]

template<>

sycl::vec< u8, 2 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u8, 2 >	min_bound,
		sycl::vec< u8, 2 >	max_bound
	)

Definition at line 154 of file mock_value.cpp.

mock_value() [56/122]

template<>

sycl::vec< u8, 3 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u8, 3 >	min_bound,
		sycl::vec< u8, 3 >	max_bound
	)

Definition at line 154 of file mock_value.cpp.

mock_value() [57/122]

template<>

sycl::vec< u8, 4 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u8, 4 >	min_bound,
		sycl::vec< u8, 4 >	max_bound
	)

Definition at line 154 of file mock_value.cpp.

mock_value() [58/122]

template<>

sycl::vec< u8, 8 > shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		sycl::vec< u8, 8 >	min_bound,
		sycl::vec< u8, 8 >	max_bound
	)

Definition at line 154 of file mock_value.cpp.

mock_value() [59/122]

template<>

u16 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		u16	min_bound,
		u16	max_bound
	)

Definition at line 28 of file mock_value.cpp.

mock_value() [60/122]

template<>

u32 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		u32	min_bound,
		u32	max_bound
	)

Definition at line 34 of file mock_value.cpp.

mock_value() [61/122]

template<>

u64 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		u64	min_bound,
		u64	max_bound
	)

Definition at line 40 of file mock_value.cpp.

mock_value() [62/122]

template<>

u8 shamalgs::primitives::mock_value	(	std::mt19937 &	eng,
		u8	min_bound,
		u8	max_bound
	)

Definition at line 22 of file mock_value.cpp.

mock_value() [63/122]

template<>

f32 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		f32	min_bound,
		f32	max_bound
	)

Definition at line 222 of file mock_value.cpp.

mock_value() [64/122]

template<>

f64 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		f64	min_bound,
		f64	max_bound
	)

Definition at line 228 of file mock_value.cpp.

mock_value() [65/122]

template<>

i16 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		i16	min_bound,
		i16	max_bound
	)

Definition at line 204 of file mock_value.cpp.

mock_value() [66/122]

template<>

i32 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		i32	min_bound,
		i32	max_bound
	)

Definition at line 210 of file mock_value.cpp.

mock_value() [67/122]

template<>

i64 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		i64	min_bound,
		i64	max_bound
	)

Definition at line 216 of file mock_value.cpp.

mock_value() [68/122]

template<>

i8 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		i8	min_bound,
		i8	max_bound
	)

Definition at line 198 of file mock_value.cpp.

mock_value() [69/122]

template<>

sycl::vec< f32, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f32, 16 >	min_bound,
		sycl::vec< f32, 16 >	max_bound
	)

Definition at line 314 of file mock_value.cpp.

mock_value() [70/122]

template<>

sycl::vec< f32, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f32, 2 >	min_bound,
		sycl::vec< f32, 2 >	max_bound
	)

Definition at line 314 of file mock_value.cpp.

mock_value() [71/122]

template<>

sycl::vec< f32, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f32, 3 >	min_bound,
		sycl::vec< f32, 3 >	max_bound
	)

Definition at line 314 of file mock_value.cpp.

mock_value() [72/122]

template<>

sycl::vec< f32, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f32, 4 >	min_bound,
		sycl::vec< f32, 4 >	max_bound
	)

Definition at line 314 of file mock_value.cpp.

mock_value() [73/122]

template<>

sycl::vec< f32, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f32, 8 >	min_bound,
		sycl::vec< f32, 8 >	max_bound
	)

Definition at line 314 of file mock_value.cpp.

mock_value() [74/122]

template<>

sycl::vec< f64, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f64, 16 >	min_bound,
		sycl::vec< f64, 16 >	max_bound
	)

Definition at line 313 of file mock_value.cpp.

mock_value() [75/122]

template<>

sycl::vec< f64, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f64, 2 >	min_bound,
		sycl::vec< f64, 2 >	max_bound
	)

Definition at line 313 of file mock_value.cpp.

mock_value() [76/122]

template<>

sycl::vec< f64, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f64, 3 >	min_bound,
		sycl::vec< f64, 3 >	max_bound
	)

Definition at line 313 of file mock_value.cpp.

mock_value() [77/122]

template<>

sycl::vec< f64, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f64, 4 >	min_bound,
		sycl::vec< f64, 4 >	max_bound
	)

Definition at line 313 of file mock_value.cpp.

mock_value() [78/122]

template<>

sycl::vec< f64, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< f64, 8 >	min_bound,
		sycl::vec< f64, 8 >	max_bound
	)

Definition at line 313 of file mock_value.cpp.

mock_value() [79/122]

template<>

sycl::vec< i16, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i16, 16 >	min_bound,
		sycl::vec< i16, 16 >	max_bound
	)

Definition at line 310 of file mock_value.cpp.

mock_value() [80/122]

template<>

sycl::vec< i16, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i16, 2 >	min_bound,
		sycl::vec< i16, 2 >	max_bound
	)

Definition at line 310 of file mock_value.cpp.

mock_value() [81/122]

template<>

sycl::vec< i16, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i16, 3 >	min_bound,
		sycl::vec< i16, 3 >	max_bound
	)

Definition at line 310 of file mock_value.cpp.

mock_value() [82/122]

template<>

sycl::vec< i16, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i16, 4 >	min_bound,
		sycl::vec< i16, 4 >	max_bound
	)

Definition at line 310 of file mock_value.cpp.

mock_value() [83/122]

template<>

sycl::vec< i16, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i16, 8 >	min_bound,
		sycl::vec< i16, 8 >	max_bound
	)

Definition at line 310 of file mock_value.cpp.

mock_value() [84/122]

template<>

sycl::vec< i32, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i32, 16 >	min_bound,
		sycl::vec< i32, 16 >	max_bound
	)

Definition at line 309 of file mock_value.cpp.

mock_value() [85/122]

template<>

sycl::vec< i32, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i32, 2 >	min_bound,
		sycl::vec< i32, 2 >	max_bound
	)

Definition at line 309 of file mock_value.cpp.

mock_value() [86/122]

template<>

sycl::vec< i32, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i32, 3 >	min_bound,
		sycl::vec< i32, 3 >	max_bound
	)

Definition at line 309 of file mock_value.cpp.

mock_value() [87/122]

template<>

sycl::vec< i32, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i32, 4 >	min_bound,
		sycl::vec< i32, 4 >	max_bound
	)

Definition at line 309 of file mock_value.cpp.

mock_value() [88/122]

template<>

sycl::vec< i32, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i32, 8 >	min_bound,
		sycl::vec< i32, 8 >	max_bound
	)

Definition at line 309 of file mock_value.cpp.

mock_value() [89/122]

template<>

sycl::vec< i64, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i64, 16 >	min_bound,
		sycl::vec< i64, 16 >	max_bound
	)

Definition at line 308 of file mock_value.cpp.

mock_value() [90/122]

template<>

sycl::vec< i64, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i64, 2 >	min_bound,
		sycl::vec< i64, 2 >	max_bound
	)

Definition at line 308 of file mock_value.cpp.

mock_value() [91/122]

template<>

sycl::vec< i64, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i64, 3 >	min_bound,
		sycl::vec< i64, 3 >	max_bound
	)

Definition at line 308 of file mock_value.cpp.

mock_value() [92/122]

template<>

sycl::vec< i64, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i64, 4 >	min_bound,
		sycl::vec< i64, 4 >	max_bound
	)

Definition at line 308 of file mock_value.cpp.

mock_value() [93/122]

template<>

sycl::vec< i64, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i64, 8 >	min_bound,
		sycl::vec< i64, 8 >	max_bound
	)

Definition at line 308 of file mock_value.cpp.

mock_value() [94/122]

template<>

sycl::vec< i8, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i8, 16 >	min_bound,
		sycl::vec< i8, 16 >	max_bound
	)

Definition at line 311 of file mock_value.cpp.

mock_value() [95/122]

template<>

sycl::vec< i8, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i8, 2 >	min_bound,
		sycl::vec< i8, 2 >	max_bound
	)

Definition at line 311 of file mock_value.cpp.

mock_value() [96/122]

template<>

sycl::vec< i8, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i8, 3 >	min_bound,
		sycl::vec< i8, 3 >	max_bound
	)

Definition at line 311 of file mock_value.cpp.

mock_value() [97/122]

template<>

sycl::vec< i8, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i8, 4 >	min_bound,
		sycl::vec< i8, 4 >	max_bound
	)

Definition at line 311 of file mock_value.cpp.

mock_value() [98/122]

template<>

sycl::vec< i8, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< i8, 8 >	min_bound,
		sycl::vec< i8, 8 >	max_bound
	)

Definition at line 311 of file mock_value.cpp.

mock_value() [99/122]

template<>

sycl::vec< u16, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u16, 16 >	min_bound,
		sycl::vec< u16, 16 >	max_bound
	)

Definition at line 305 of file mock_value.cpp.

mock_value() [100/122]

template<>

sycl::vec< u16, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u16, 2 >	min_bound,
		sycl::vec< u16, 2 >	max_bound
	)

Definition at line 305 of file mock_value.cpp.

mock_value() [101/122]

template<>

sycl::vec< u16, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u16, 3 >	min_bound,
		sycl::vec< u16, 3 >	max_bound
	)

Definition at line 305 of file mock_value.cpp.

mock_value() [102/122]

template<>

sycl::vec< u16, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u16, 4 >	min_bound,
		sycl::vec< u16, 4 >	max_bound
	)

Definition at line 305 of file mock_value.cpp.

mock_value() [103/122]

template<>

sycl::vec< u16, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u16, 8 >	min_bound,
		sycl::vec< u16, 8 >	max_bound
	)

Definition at line 305 of file mock_value.cpp.

mock_value() [104/122]

template<>

sycl::vec< u32, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u32, 16 >	min_bound,
		sycl::vec< u32, 16 >	max_bound
	)

Definition at line 304 of file mock_value.cpp.

mock_value() [105/122]

template<>

sycl::vec< u32, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u32, 2 >	min_bound,
		sycl::vec< u32, 2 >	max_bound
	)

Definition at line 304 of file mock_value.cpp.

mock_value() [106/122]

template<>

sycl::vec< u32, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u32, 3 >	min_bound,
		sycl::vec< u32, 3 >	max_bound
	)

Definition at line 304 of file mock_value.cpp.

mock_value() [107/122]

template<>

sycl::vec< u32, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u32, 4 >	min_bound,
		sycl::vec< u32, 4 >	max_bound
	)

Definition at line 304 of file mock_value.cpp.

mock_value() [108/122]

template<>

sycl::vec< u32, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u32, 8 >	min_bound,
		sycl::vec< u32, 8 >	max_bound
	)

Definition at line 304 of file mock_value.cpp.

mock_value() [109/122]

template<>

sycl::vec< u64, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u64, 16 >	min_bound,
		sycl::vec< u64, 16 >	max_bound
	)

Definition at line 303 of file mock_value.cpp.

mock_value() [110/122]

template<>

sycl::vec< u64, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u64, 2 >	min_bound,
		sycl::vec< u64, 2 >	max_bound
	)

Definition at line 303 of file mock_value.cpp.

mock_value() [111/122]

template<>

sycl::vec< u64, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u64, 3 >	min_bound,
		sycl::vec< u64, 3 >	max_bound
	)

Definition at line 303 of file mock_value.cpp.

mock_value() [112/122]

template<>

sycl::vec< u64, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u64, 4 >	min_bound,
		sycl::vec< u64, 4 >	max_bound
	)

Definition at line 303 of file mock_value.cpp.

mock_value() [113/122]

template<>

sycl::vec< u64, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u64, 8 >	min_bound,
		sycl::vec< u64, 8 >	max_bound
	)

Definition at line 303 of file mock_value.cpp.

mock_value() [114/122]

template<>

sycl::vec< u8, 16 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u8, 16 >	min_bound,
		sycl::vec< u8, 16 >	max_bound
	)

Definition at line 306 of file mock_value.cpp.

mock_value() [115/122]

template<>

sycl::vec< u8, 2 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u8, 2 >	min_bound,
		sycl::vec< u8, 2 >	max_bound
	)

Definition at line 306 of file mock_value.cpp.

mock_value() [116/122]

template<>

sycl::vec< u8, 3 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u8, 3 >	min_bound,
		sycl::vec< u8, 3 >	max_bound
	)

Definition at line 306 of file mock_value.cpp.

mock_value() [117/122]

template<>

sycl::vec< u8, 4 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u8, 4 >	min_bound,
		sycl::vec< u8, 4 >	max_bound
	)

Definition at line 306 of file mock_value.cpp.

mock_value() [118/122]

template<>

sycl::vec< u8, 8 > shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		sycl::vec< u8, 8 >	min_bound,
		sycl::vec< u8, 8 >	max_bound
	)

Definition at line 306 of file mock_value.cpp.

mock_value() [119/122]

template<>

u16 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		u16	min_bound,
		u16	max_bound
	)

Definition at line 180 of file mock_value.cpp.

mock_value() [120/122]

template<>

u32 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		u32	min_bound,
		u32	max_bound
	)

Definition at line 186 of file mock_value.cpp.

mock_value() [121/122]

template<>

u64 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		u64	min_bound,
		u64	max_bound
	)

Definition at line 192 of file mock_value.cpp.

mock_value() [122/122]

template<>

u8 shamalgs::primitives::mock_value	(	std::mt19937_64 &	eng,
		u8	min_bound,
		u8	max_bound
	)

Definition at line 174 of file mock_value.cpp.

mock_vector() [1/2]

template<class T >

std::vector< T > shamalgs::primitives::mock_vector ( u64  seed, u32  len  )

inline

Generates a vector of random mock values using default bounds.

This function generates a std::vector of random values of type T using the default minimum and maximum bounds defined by shambase::VectorProperties<T>. This is a convenience function that automatically determines appropriate bounds based on the type T. The function uses a deterministic random number generator seeded with the provided seed value.

Template Parameters

T	The type of values to generate (supports primitive types and SYCL vectors)

Parameters

seed	The seed value for the random number generator
len	The length of the vector to generate

Returns

A std::vector<T> containing len random values within the default bounds for the type

#include "shamalgs/primitives/mock_vector.hpp"
 
// Generate vectors with default bounds
std::vector<i32> random_ints = shamalgs::mock_vector<i32>(42, 10);
std::vector<f64> random_doubles = shamalgs::mock_vector<f64>(123, 5);
std::vector<sycl::vec<f32, 2>> random_vecs = shamalgs::mock_vector<sycl::vec<f32, 2>>(456,
3);

Definition at line 100 of file mock_vector.hpp.

Here is the call graph for this function:

mock_vector() [2/2]

template<class T >

std::vector< T > shamalgs::primitives::mock_vector	(	u64	seed,
		u32	len,
		T	min_bound,
		T	max_bound
	)

Generates a vector of random mock values within specified bounds.

This function generates a std::vector of random values of type T within the range [min_bound, max_bound]. Each element in the vector is generated independently using the same bounds. The function uses a deterministic random number generator seeded with the provided seed value, ensuring reproducible results for the same seed.

Template Parameters

T	The type of values to generate (supports primitive types and SYCL vectors)

Parameters

seed	The seed value for the random number generator
len	The length of the vector to generate
min_bound	The minimum bound for the generated values
max_bound	The maximum bound for the generated values

Returns

A std::vector<T> containing len random values within the specified bounds

#include "shamalgs/primitives/mock_vector.hpp"
 
// Generate a vector of random integers
std::vector<i32> random_ints = shamalgs::mock_vector<i32>(42, 10, i32{0}, i32{100});
 
// Generate a vector of random floats
std::vector<f32> random_floats = shamalgs::mock_vector<f32>(123, 5, f32{0.0f}, f32{1.0f});
 
// Generate a vector of random SYCL vectors
std::vector<sycl::vec<f64, 3>> random_vecs = shamalgs::mock_vector<sycl::vec<f64, 3>>(
    456,
    3,
    sycl::vec<f64, 3>{0.0, 0.0, 0.0},
    sycl::vec<f64, 3>{1.0, 1.0, 1.0}
);

Definition at line 62 of file mock_vector.hpp.

Here is the call graph for this function:

reduction_impl_from_params()

REDUCTION_IMPL shamalgs::primitives::reduction_impl_from_params ( const std::string &  impl )

inline

Definition at line 50 of file reduction.cpp.

reduction_impl_to_params()

shamalgs::impl_param shamalgs::primitives::reduction_impl_to_params ( const REDUCTION_IMPL &  impl )

inline

Definition at line 68 of file reduction.cpp.

scan_exclusive_sum_in_place()

template<class T >

void shamalgs::primitives::scan_exclusive_sum_in_place	(	sham::DeviceBuffer< T > &	buf1,
		u32	len
	)

Compute exclusive prefix sum in-place on a device buffer.

Performs an exclusive scan (prefix sum) operation on the device buffer, modifying it in-place. For each position i, the result contains the sum of all elements from index 0 to i-1. The first element is set to 0 (the identity for addition).

The operation transforms: [a, b, c, d] -> [0, a, a+b, a+b+c]

Template Parameters

T	Element type - must support addition and assignment

Parameters

buf1	Device buffer to scan in-place (modified by this operation)
len	Number of elements to process from the beginning of the buffer

Note

The buffer is modified in-place; original values are overwritten

// Example: Basic exclusive scan
std::vector<u32> data = {1, 2, 3, 4, 5};
sham::DeviceBuffer<u32> buffer(data);
 
// Perform in-place exclusive scan
scan_exclusive_sum_in_place(buffer, data.size());
 
// Buffer now contains: [0, 1, 3, 6, 10]
auto result = buffer.copy_to_stdvec();
// result[0] = 0, result[1] = 1, result[2] = 3, result[3] = 6, result[4] = 10
 
// Example: Partial scan
std::vector<u32> values = {5, 10, 15, 20, 25, 30};
sham::DeviceBuffer<u32> partial_buffer(values);
 
// Scan only first 4 elements
scan_exclusive_sum_in_place(partial_buffer, 4);
 
// Buffer now contains: [0, 5, 15, 30, 25, 30]
// Only first 4 elements were modified

Definition at line 212 of file scan_exclusive_sum_in_place.cpp.

Here is the call graph for this function:

scan_exclusive_sum_in_place_impl_from_params()

EXSCAN_IN_PLACE_IMPL shamalgs::primitives::scan_exclusive_sum_in_place_impl_from_params ( const std::string &  impl )

inline

Definition at line 139 of file scan_exclusive_sum_in_place.cpp.

scan_exclusive_sum_in_place_impl_to_params()

shamalgs::impl_param shamalgs::primitives::scan_exclusive_sum_in_place_impl_to_params ( const EXSCAN_IN_PLACE_IMPL &  impl )

inline

Definition at line 163 of file scan_exclusive_sum_in_place.cpp.

segmented_sort_in_place< u32 >()

template<>

void shamalgs::primitives::segmented_sort_in_place< u32 >	(	sham::DeviceBuffer< u32 > &	buf,
		const sham::DeviceBuffer< u32 > &	offsets
	)

Definition at line 203 of file segmented_sort_in_place.cpp.

segmented_sort_in_place< u32_2 >()

template<>

void shamalgs::primitives::segmented_sort_in_place< u32_2 >	(	sham::DeviceBuffer< u32_2 > &	buf,
		const sham::DeviceBuffer< u32 > &	offsets
	)

Definition at line 194 of file segmented_sort_in_place.cpp.

sort_by_key() [1/2]

template<class Tkey , class Tval >

void shamalgs::primitives::sort_by_key	(	const sham::DeviceScheduler_ptr &	sched,
		sham::DeviceBuffer< Tkey > &	buf_key,
		sham::DeviceBuffer< Tval > &	buf_values,
		u32	len
	)

Sort key-value pairs using USM buffers.

Performs an in-place parallel sort of key-value pairs where the values are reordered according to the sorted order of their corresponding keys.

Template Parameters

Tkey	Key type - must be comparable (supports < operator)
Tval	Value type - can be any copyable type

Parameters

sched	sham::DeviceScheduler_ptr for execution
buf_key	Device buffer containing the keys to sort by
buf_values	Device buffer containing the values to reorder
len	Length of both buffers

Exceptions

std::invalid_argument

if len is not a power of 2

Note

The function modifies both buffers in-place

This function currently only supports powers of 2

// Example: Sort data by keys using USM buffers
auto sched = shamsys::instance::get_compute_scheduler_ptr();
sham::DeviceBuffer<double> keys(input_keys, N);
sham::DeviceBuffer<DataType> values(input_values, N);
 
// Sort values according to key order
sort_by_key(sched, keys, values, N);

Definition at line 159 of file sort_by_keys.hpp.

Here is the call graph for this function:

sort_by_key() [2/2]

template<class Tkey , class Tval >

void shamalgs::primitives::sort_by_key	(	sycl::queue &	q,
		sycl::buffer< Tkey > &	buf_key,
		sycl::buffer< Tval > &	buf_values,
		u32	len
	)

Sort key-value pairs using sycl::buffers.

Performs an in-place parallel sort of key-value pairs where the values are reordered according to the sorted order of their corresponding keys.

Template Parameters

Tkey	Key type - must be comparable (supports < operator)
Tval	Value type - can be any copyable type

Parameters

q	sycl::queue for device execution
buf_key	Buffer containing the keys to sort by
buf_values	Buffer containing the values to reorder
len	Length of both buffers

Exceptions

std::invalid_argument

if len is not a power of 2

Note

The function modifies both buffers in-place

This function currently only supports powers of 2

// Example: Sort data by keys
sycl::queue q;
sycl::buffer<double> keys(input_keys, N);
sycl::buffer<DataType> values(input_values, N);
 
// Sort values according to key order
sort_by_key(q, keys, values, N);

Definition at line 123 of file sort_by_keys.hpp.

Here is the call graph for this function:

sort_by_key_pow2_len() [1/2]

template<class Tkey , class Tval >

void shamalgs::primitives::sort_by_key_pow2_len	(	const sham::DeviceScheduler_ptr &	sched,
		sham::DeviceBuffer< Tkey > &	buf_key,
		sham::DeviceBuffer< Tval > &	buf_values,
		u32	len
	)

Sort key-value pairs using USM buffers (power-of-2 optimized)

Performs an in-place parallel sort of key-value pairs where the values are reordered according to the sorted order of their corresponding keys.

Template Parameters

Tkey	Key type - must be comparable (supports < operator)
Tval	Value type - can be any copyable type

Parameters

sched	sham::DeviceScheduler_ptr for execution
buf_key	Device buffer containing the keys to sort by
buf_values	Device buffer containing the values to reorder
len	Length of both buffers (must be a power of 2)

Note

The function modifies both buffers in-place

// Example: Sort data by keys using USM buffers
auto sched = shamsys::instance::get_compute_scheduler_ptr();
sham::DeviceBuffer<float> keys(input_keys, N);
sham::DeviceBuffer<DataType> values(input_values, N);
 
// Sort values according to key order
sort_by_key_pow2_len(sched, keys, values, N);

Definition at line 36 of file sort_by_keys.cpp.

sort_by_key_pow2_len() [2/2]

template<class Tkey , class Tval >

void shamalgs::primitives::sort_by_key_pow2_len	(	sycl::queue &	q,
		sycl::buffer< Tkey > &	buf_key,
		sycl::buffer< Tval > &	buf_values,
		u32	len
	)

Sort key-value pairs using sycl::buffers (power-of-2 optimized)

Performs an in-place parallel sort of key-value pairs where the values are reordered according to the sorted order of their corresponding keys.

Template Parameters

Tkey	Key type - must be comparable (supports < operator)
Tval	Value type - can be any copyable type

Parameters

q	sycl::queue for device execution
buf_key	Buffer containing the keys to sort by
buf_values	Buffer containing the values to reorder
len	Length of both buffers (must be a power of 2)

Note

The function modifies both buffers in-place

// Example: Sort data by keys
sycl::queue q;
sycl::buffer<float> keys(input_keys, N);
sycl::buffer<DataType> values(input_values, N);
 
// Sort values according to key order
sort_by_key_pow2_len(q, keys, values, N);

Definition at line 24 of file sort_by_keys.cpp.

sum()

template<class T >

T shamalgs::primitives::sum	(	const sham::DeviceScheduler_ptr &	sched,
		const sham::DeviceBuffer< T > &	buf1,
		u32	start_id,
		u32	end_id
	)

Compute the sum of elements in a device buffer within a specified range.

This function computes the sum of all elements in the buffer between start_id (inclusive) and end_id (exclusive). The computation is performed on the device using the provided scheduler.

Template Parameters

T	The data type of elements in the buffer (e.g., float, double, int).

Parameters

sched	The device scheduler to run on.
buf1	The input buffer containing the elements to sum.
start_id	The starting index (inclusive) of the range to sum.
end_id	The ending index (exclusive) of the range to sum.

Returns

T The computed sum of elements in the specified range.

Precondition

start_id <= end_id

end_id <= buf1.get_size()

Example:

auto sched = shamsys::get_compute_Scheduler_ptr();
 
sham::DeviceBuffer<double> values = ...;
u32 start = 0;
u32 end = values.get_size();
 
double total = shamalgs::primitives::sum(sched, values, start, end);

values = {1.0, 2.0, 3.0, 4.0, 5.0}, start = 1, end = 4 result = 9.0 (2.0 + 3.0 + 4.0)

Definition at line 105 of file reduction.cpp.

Here is the call graph for this function:

unflatten_buffer()

template<class Tvec , sham::USMKindTarget target>

sham::DeviceBuffer< Tvec, target > shamalgs::primitives::unflatten_buffer ( const sham::DeviceBuffer< typename shambase::VectorProperties< Tvec >::component_type, target > &  buffer )

inline

Unflatten a buffer that contains a flattened vector.

Parameters

buffer

The buffer to unflatten

Returns

A new buffer of type Tvec with the same size as the original buffer divided by the number of components in the vector.

Exceptions

std::invalid_argument	if the buffer has a size that is not a multiple of the number of components in the vector.
std::runtime_error	if the buffer is not a device buffer.

Definition at line 116 of file flatten.hpp.

Here is the call graph for this function:

Variable Documentation

is_all_true_impl

IS_ALL_TRUE_IMPL shamalgs::primitives::is_all_true_impl = IS_ALL_TRUE_IMPL::HOST

Definition at line 70 of file is_all_true.cpp.

reduction_impl

REDUCTION_IMPL shamalgs::primitives::reduction_impl = get_default_reduction_impl()

Definition at line 48 of file reduction.cpp.

scan_exclusive_sum_in_place_impl

EXSCAN_IN_PLACE_IMPL shamalgs::primitives::scan_exclusive_sum_in_place_impl = get_default_scan_exclusive_sum_in_place_impl()

Definition at line 136 of file scan_exclusive_sum_in_place.cpp.