KernelOps< T, Bits, Arch > Struct Template Reference

#include <kernel_ops.h>

Public Types
using	K = Microkernel< T, Bits, Arch >

Static Public Member Functions
template<typename Expr >
static FORCE_INLINE void	eval (T *output, const Expr &expr) noexcept
	Evaluation: Dispatch: pick contiguous or permuted eval based on expression layout.
template<typename Expr >
static FORCE_INLINE T	reduce_min (const Expr &expr) noexcept
template<typename Expr >
static FORCE_INLINE T	reduce_max (const Expr &expr) noexcept
template<typename Expr >
static FORCE_INLINE T	reduce_sum (const Expr &expr) noexcept
template<typename Expr1 , typename Expr2 >
static FORCE_INLINE bool	reduce_all_approx_equal (const Expr1 &lhs, const Expr2 &rhs, T tolerance) noexcept
	Check if all logical elements of two expressions are approximately equal.
template<typename Expr1 , typename Expr2 >
static FORCE_INLINE T	dot (const Expr1 &expr1, my_size_t base1, my_size_t stride1, const Expr2 &expr2, my_size_t base2, my_size_t stride2, my_size_t len) noexcept
	Dispatch dot product based on stride values.
template<typename Expr1 , typename Expr2 >
static FORCE_INLINE T	naive_dot_physical (const Expr1 &expr1, my_size_t base1, my_size_t stride1, const Expr2 &expr2, my_size_t base2, my_size_t stride2, my_size_t len) noexcept
	Naive scalar dot product for testing/validation.

Static Public Attributes
static constexpr my_size_t	simdWidth = K::simdWidth

Member Typedef Documentation

K

template<typename T , my_size_t Bits, typename Arch >

using KernelOps< T, Bits, Arch >::K = Microkernel<T, Bits, Arch>

Member Function Documentation

dot()

template<typename T , my_size_t Bits, typename Arch >

template<typename Expr1 , typename Expr2 >

static FORCE_INLINE T KernelOps< T, Bits, Arch >::dot ( const Expr1 &  expr1, my_size_t  base1, my_size_t  stride1, const Expr2 &  expr2, my_size_t  base2, my_size_t  stride2, my_size_t  len  )

inlinestaticnoexcept

Dispatch dot product based on stride values.

Here is the call graph for this function:

eval()

template<typename T , my_size_t Bits, typename Arch >

template<typename Expr >

static FORCE_INLINE void KernelOps< T, Bits, Arch >::eval ( T *  output, const Expr &  expr  )

inlinestaticnoexcept

Evaluation: Dispatch: pick contiguous or permuted eval based on expression layout.

Here is the call graph for this function:

naive_dot_physical()

template<typename T , my_size_t Bits, typename Arch >

template<typename Expr1 , typename Expr2 >

static FORCE_INLINE T KernelOps< T, Bits, Arch >::naive_dot_physical ( const Expr1 &  expr1, my_size_t  base1, my_size_t  stride1, const Expr2 &  expr2, my_size_t  base2, my_size_t  stride2, my_size_t  len  )

inlinestaticnoexcept

Naive scalar dot product for testing/validation.

Here is the call graph for this function:

reduce_all_approx_equal()

template<typename T , my_size_t Bits, typename Arch >

template<typename Expr1 , typename Expr2 >

static FORCE_INLINE bool KernelOps< T, Bits, Arch >::reduce_all_approx_equal ( const Expr1 &  lhs, const Expr2 &  rhs, T  tolerance  )

inlinestaticnoexcept

Check if all logical elements of two expressions are approximately equal.

Here is the call graph for this function:

reduce_max()

template<typename T , my_size_t Bits, typename Arch >

template<typename Expr >

static FORCE_INLINE T KernelOps< T, Bits, Arch >::reduce_max ( const Expr &  expr )

inlinestaticnoexcept

Here is the call graph for this function:

reduce_min()

template<typename T , my_size_t Bits, typename Arch >

template<typename Expr >

static FORCE_INLINE T KernelOps< T, Bits, Arch >::reduce_min ( const Expr &  expr )

inlinestaticnoexcept

Here is the call graph for this function:

reduce_sum()

template<typename T , my_size_t Bits, typename Arch >

template<typename Expr >

static FORCE_INLINE T KernelOps< T, Bits, Arch >::reduce_sum ( const Expr &  expr )

inlinestaticnoexcept

Here is the call graph for this function:

Member Data Documentation

simdWidth

template<typename T , my_size_t Bits, typename Arch >

constexpr my_size_t KernelOps< T, Bits, Arch >::simdWidth = K::simdWidth

staticconstexpr

---

The documentation for this struct was generated from the following file:

• core/include/fused/kernel_ops/kernel_ops.h