Load

Generate Address

Represents GEN_ADDR instruction.

Intrinsics Prototype and Arguments

Intrinsics Definition

__global void* gen_addr(int5 inx, int8_t tensor, int switches=0, __global void *income={}, bool predicate=1, bool polarity=0)

Parameters

inx	Tensor coordinates (SRC1).
tensor	Tensor number.
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

Pointer in global address space, pointing to the selected tensor element.

Load local

Represents LD_L instruction.

Loads a scalar value from MMIO or SLM(Scalar Local Memory) into a SRF.

Intrinsics Prototype and Arguments

Intrinsics Definition

float s_f32_ld_l(uint32_t addr, int switches=0, float income={}, bool predicate=1, bool polarity=0)

bf16 s_bf16_ld_l(uint32_t addr, int switches=0, bf16 income={}, bool predicate=1, bool polarity=0)

int32_t s_i32_ld_l(uint32_t addr, int switches=0, int32_t income={}, bool predicate=1, bool polarity=0)

uint32_t s_u32_ld_l(uint32_t addr, int switches=0, uint32_t income={}, bool predicate=1, bool polarity=0)

int16_t s_i16_ld_l(uint32_t addr, int switches=0, int16_t income={}, bool predicate=1, bool polarity=0)

uint16_t s_u16_ld_l(uint32_t addr, int switches=0, uint16_t income={}, bool predicate=1, bool polarity=0)

int8_t s_i8_ld_l(uint32_t addr, int switches=0, int8_t income={}, bool predicate=1, bool polarity=0)

uint8_t s_u8_ld_l(uint32_t addr, int switches=0, uint8_t income={}, bool predicate=1, bool polarity=0)

bool s_i1_ld_l(uint32_t addr, int switches=0, bool income={}, bool predicate=1, bool polarity=0)

Parameters

addr	Address to read from (SRC1).
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Switches

Allowed switches are:

• SW_MMIO - load from MMIO.

• SW_SLM - load from SLM (the default).

Returns

The loaded scalar value from local memory.

Load local vector

Represents LD_L_V instruction.

Loads a vector from VLM(Vector Local Memory) into a VRF

Intrinsics Prototype and Arguments

Intrinsics Definition

float64 v_f32_ld_l_v_b(uint32_t addr, int switches=0, float64 income={}, bool predicate=1, bool polarity=0)

bfloat128 v_bf16_ld_l_v_vb(uint32_t addr, int switches, bfloat128 income, bool128 predicate, bool polarity=0)

bfloat128 v_bf16_ld_l_v_b(uint32_t addr, int switches=0, bfloat128 income={}, bool predicate=1, bool polarity=0)

int64 v_i32_ld_l_v_b(uint32_t addr, int switches=0, int64 income={}, bool predicate=1, bool polarity=0)

uint64 v_u32_ld_l_v_b(uint32_t addr, int switches=0, uint64 income={}, bool predicate=1, bool polarity=0)

short128 v_i16_ld_l_v_b(uint32_t addr, int switches=0, short128 income={}, bool predicate=1, bool polarity=0)

ushort128 v_u16_ld_l_v_b(uint32_t addr, int switches=0, ushort128 income={}, bool predicate=1, bool polarity=0)

char256 v_i8_ld_l_v_b(uint32_t addr, int switches=0, char256 income={}, bool predicate=1, bool polarity=0)

uchar256 v_u8_ld_l_v_b(uint32_t addr, int switches=0, uchar256 income={}, bool predicate=1, bool polarity=0)

bool256 v_i1_ld_l_v_b(uint32_t addr, int switches=0, bool256 income={}, bool predicate=1, bool polarity=0)

float64 v_f32_ld_l_v_vb(uint32_t addr, int switches, float64 income, bool64 predicate, bool polarity=0)

int64 v_i32_ld_l_v_vb(uint32_t addr, int switches, int64 income, bool64 predicate, bool polarity=0)

uint64 v_u32_ld_l_v_vb(uint32_t addr, int switches, uint64 income, bool64 predicate, bool polarity=0)

short128 v_i16_ld_l_v_vb(uint32_t addr, int switches, short128 income, bool128 predicate, bool polarity=0)

ushort128 v_u16_ld_l_v_vb(uint32_t addr, int switches, ushort128 income, bool128 predicate, bool polarity=0)

char256 v_i8_ld_l_v_vb(uint32_t addr, int switches, char256 income, bool256 predicate, bool polarity=0)

uchar256 v_u8_ld_l_v_vb(uint32_t addr, int switches, uchar256 income, bool256 predicate, bool polarity=0)

bool256 v_i1_ld_l_v_vb(uint32_t addr, int switches, bool256 income, bool256 predicate, bool polarity=0)

Parameters

addr	Address to read from (SRC1).
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

The loaded vector value from local memory.

Load local vector high

Represents LD_L_V_HIGH instruction.

Intrinsics Prototype and Arguments

Intrinsics Definition

float64 v_f32_ld_l_v_high_vb(uint32_t addr, int switches, float64 income, bool64 predicate, bool polarity=0)

float64 v_f32_ld_l_v_high_b(uint32_t addr, int switches=0, float64 income={}, bool predicate=1, bool polarity=0)

bfloat128 v_bf16_ld_l_v_high_vb(uint32_t addr, int switches, bfloat128 income, bool128 predicate, bool polarity=0)

bfloat128 v_bf16_ld_l_v_high_b(uint32_t addr, int switches=0, bfloat128 income={}, bool predicate=1, bool polarity=0)

int64 v_i32_ld_l_v_high_vb(uint32_t addr, int switches, int64 income, bool64 predicate, bool polarity=0)

int64 v_i32_ld_l_v_high_b(uint32_t addr, int switches=0, int64 income={}, bool predicate=1, bool polarity=0)

uint64 v_u32_ld_l_v_high_vb(uint32_t addr, int switches, uint64 income, bool64 predicate, bool polarity=0)

uint64 v_u32_ld_l_v_high_b(uint32_t addr, int switches=0, uint64 income={}, bool predicate=1, bool polarity=0)

short128 v_i16_ld_l_v_high_vb(uint32_t addr, int switches, short128 income, bool128 predicate, bool polarity=0)

short128 v_i16_ld_l_v_high_b(uint32_t addr, int switches=0, short128 income={}, bool predicate=1, bool polarity=0)

ushort128 v_u16_ld_l_v_high_vb(uint32_t addr, int switches, ushort128 income, bool128 predicate, bool polarity=0)

ushort128 v_u16_ld_l_v_high_b(uint32_t addr, int switches=0, ushort128 income={}, bool predicate=1, bool polarity=0)

char256 v_i8_ld_l_v_high_vb(uint32_t addr, int switches, char256 income, bool256 predicate, bool polarity=0)

char256 v_i8_ld_l_v_high_b(uint32_t addr, int switches=0, char256 income={}, bool predicate=1, bool polarity=0)

uchar256 v_u8_ld_l_v_high_vb(uint32_t addr, int switches, uchar256 income, bool256 predicate, bool polarity=0)

uchar256 v_u8_ld_l_v_high_b(uint32_t addr, int switches=0, uchar256 income={}, bool predicate=1, bool polarity=0)

bool256 v_i1_ld_l_v_high_vb(uint32_t addr, int switches, bool256 income, bool256 predicate, bool polarity=0)

bool256 v_i1_ld_l_v_high_b(uint32_t addr, int switches=0, bool256 income={}, bool predicate=1, bool polarity=0)

Parameters

addr	Address to read from (SRC1).
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

The loaded half vector value from local memory.

Load local vector low

Represents LD_L_V_LOW instruction.

Intrinsics Prototype and Arguments

Intrinsics Definition

float64 v_f32_ld_l_v_low_vb(uint32_t addr, int switches, float64 income, bool64 predicate, bool polarity=0)

float64 v_f32_ld_l_v_low_b(uint32_t addr, int switches=0, float64 income={}, bool predicate=1, bool polarity=0)

bfloat128 v_bf16_ld_l_v_low_vb(uint32_t addr, int switches, bfloat128 income, bool128 predicate, bool polarity=0)

bfloat128 v_bf16_ld_l_v_low_b(uint32_t addr, int switches=0, bfloat128 income={}, bool predicate=1, bool polarity=0)

int64 v_i32_ld_l_v_low_vb(uint32_t addr, int switches, int64 income, bool64 predicate, bool polarity=0)

int64 v_i32_ld_l_v_low_b(uint32_t addr, int switches=0, int64 income={}, bool predicate=1, bool polarity=0)

uint64 v_u32_ld_l_v_low_vb(uint32_t addr, int switches, uint64 income, bool64 predicate, bool polarity=0)

uint64 v_u32_ld_l_v_low_b(uint32_t addr, int switches=0, uint64 income={}, bool predicate=1, bool polarity=0)

short128 v_i16_ld_l_v_low_vb(uint32_t addr, int switches, short128 income, bool128 predicate, bool polarity=0)

short128 v_i16_ld_l_v_low_b(uint32_t addr, int switches=0, short128 income={}, bool predicate=1, bool polarity=0)

ushort128 v_u16_ld_l_v_low_vb(uint32_t addr, int switches, ushort128 income, bool128 predicate, bool polarity=0)

ushort128 v_u16_ld_l_v_low_b(uint32_t addr, int switches=0, ushort128 income={}, bool predicate=1, bool polarity=0)

char256 v_i8_ld_l_v_low_vb(uint32_t addr, int switches, char256 income, bool256 predicate, bool polarity=0)

char256 v_i8_ld_l_v_low_b(uint32_t addr, int switches=0, char256 income={}, bool predicate=1, bool polarity=0)

uchar256 v_u8_ld_l_v_low_vb(uint32_t addr, int switches, uchar256 income, bool256 predicate, bool polarity=0)

uchar256 v_u8_ld_l_v_low_b(uint32_t addr, int switches=0, uchar256 income={}, bool predicate=1, bool polarity=0)

bool256 v_i1_ld_l_v_low_vb(uint32_t addr, int switches, bool256 income, bool256 predicate, bool polarity=0)

bool256 v_i1_ld_l_v_low_b(uint32_t addr, int switches=0, bool256 income={}, bool predicate=1, bool polarity=0)

Parameters

addr	Address to read from (SRC1).
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

The loaded half vector value from local memory.

Load Global

Represents LD_G instruction. Loads a single scalar value from global memory, if the destination is a vector register, the value is being replicated across all lanes.

Load a value from the given address. If the value is stored into a VRF then the same value is broadcasted to the entire vector register.

Intrinsics Prototype and Arguments

Intrinsics Definition

float s_f32_ld_g(__global void *addr, int switches=0, float income={}, bool predicate=1, bool polarity=0)

bf16 s_bf16_ld_g(__global void *addr, int switches=0, bf16 income={}, bool predicate=1, bool polarity=0)

int32_t s_i32_ld_g(__global void *addr, int switches=0, int32_t income={}, bool predicate=1, bool polarity=0)

uint32_t s_u32_ld_g(__global void *addr, int switches=0, uint32_t income={}, bool predicate=1, bool polarity=0)

int16_t s_i16_ld_g(__global void *addr, int switches=0, int16_t income={}, bool predicate=1, bool polarity=0)

uint16_t s_u16_ld_g(__global void *addr, int switches=0, uint16_t income={}, bool predicate=1, bool polarity=0)

int8_t s_i8_ld_g(__global void *addr, int switches=0, int8_t income={}, bool predicate=1, bool polarity=0)

uint8_t s_u8_ld_g(__global void *addr, int switches=0, uint8_t income={}, bool predicate=1, bool polarity=0)

bool s_i1_ld_g(__global void *addr, int switches=0, bool income={}, bool predicate=1, bool polarity=0)

bfloat128 v_bf16_ld_g(__global void *addr, int switches=0, bfloat128 income={}, bool predicate=1, bool polarity=0)

float64 v_f32_ld_g(__global void *addr, int switches=0, float64 income={}, bool predicate=1, bool polarity=0)

int64 v_i32_ld_g(__global void *addr, int switches=0, int64 income={}, bool predicate=1, bool polarity=0)

uint64 v_u32_ld_g(__global void *addr, int switches=0, uint64 income={}, bool predicate=1, bool polarity=0)

short128 v_i16_ld_g(__global void *addr, int switches=0, short128 income={}, bool predicate=1, bool polarity=0)

ushort128 v_u16_ld_g(__global void *addr, int switches=0, ushort128 income={}, bool predicate=1, bool polarity=0)

char256 v_i8_ld_g(__global void *addr, int switches=0, char256 income={}, bool predicate=1, bool polarity=0)

uchar256 v_u8_ld_g(__global void *addr, int switches=0, uchar256 income={}, bool predicate=1, bool polarity=0)

Parameters

addr	Address to read from (SRC1).
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

The loaded scalar value from global memory.

Load Tensor

Represents LD_TNSR instruction.

Loads a vector register with the values from the given tensor based on specified index values.

Intrinsics Prototype and Arguments

Intrinsics Definition

float64 v_f32_ld_tnsr_b(int5 ndx, const int8_t tensor, int switches=0, float64 income={}, bool predicate=1, bool polarity=0)

int64 v_i32_ld_tnsr_b(int5 ndx, const int8_t tensor, int switches=0, int64 income={}, bool predicate=1, bool polarity=0)

uint64 v_u32_ld_tnsr_b(int5 ndx, const int8_t tensor, int switches=0, uint64 income={}, bool predicate=1, bool polarity=0)

short128 v_i16_ld_tnsr_b(int5 ndx, const int8_t tensor, int switches=0, short128 income={}, bool predicate=1, bool polarity=0)

ushort128 v_u16_ld_tnsr_b(int5 ndx, const int8_t tensor, int switches=0, ushort128 income={}, bool predicate=1, bool polarity=0)

char256 v_i8_ld_tnsr_b(int5 ndx, const int8_t tensor, int switches=0, char256 income={}, bool predicate=1, bool polarity=0)

uchar256 v_u8_ld_tnsr_b(int5 ndx, const int8_t tensor, int switches=0, uchar256 income={}, bool predicate=1, bool polarity=0)

bool256 v_i1_ld_tnsr_b(int5 ndx, const int8_t tensor, int switches=0, bool256 income={}, bool predicate=1, bool polarity=0)

float64 v_f32_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, float64 income, bool64 predicate, bool polarity=0)

bfloat128 v_bf16_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, bfloat128 income, bool128 predicate, bool polarity=0)

int64 v_i32_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, int64 income, bool64 predicate, bool polarity=0)

uint64 v_u32_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, uint64 income, bool64 predicate, bool polarity=0)

short128 v_i16_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, short128 income, bool128 predicate, bool polarity=0)

ushort128 v_u16_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, ushort128 income, bool128 predicate, bool polarity=0)

char256 v_i8_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, char256 income, bool256 predicate, bool polarity=0)

uchar256 v_u8_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, uchar256 income, bool256 predicate, bool polarity=0)

bool256 v_i1_ld_tnsr_vb(int5 ndx, int8_t tensor, int switches, bool256 income, bool256 predicate, bool polarity=0)

bfloat128 v_bf16_ld_tnsr_b(int5 ndx, int8_t tensor, int switches=0, bfloat128 income={}, bool predicate=1, bool polarity=0)

Parameters

ndx	Tensor coordinates (SRC1).
tensor	Tensor number.
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

The loaded values.

Load tensor using PARTIAL switch

Represents LD_TNSR instruction with PARTIAL switch. Loads elements from global memory into a subset of vector lanes.

Intrinsics Prototype and Arguments

Intrinsics Definition

float64 v_f32_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, float64 income={}, bool predicate=1, bool polarity=0)

bfloat128 v_bf16_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, bfloat128 income={}, bool predicate=1, bool polarity=0)

int64 v_i32_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, int64 income={}, bool predicate=1, bool polarity=0)

uint64 v_u32_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, uint64 income={}, bool predicate=1, bool polarity=0)

short128 v_i16_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, short128 income={}, bool predicate=1, bool polarity=0)

ushort128 v_u16_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, ushort128 income={}, bool predicate=1, bool polarity=0)

char256 v_i8_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, char256 income={}, bool predicate=1, bool polarity=0)

uchar256 v_u8_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, uchar256 income={}, bool predicate=1, bool polarity=0)

bool256 v_i1_ld_tnsr_partial_b(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches=0, bool256 income={}, bool predicate=1, bool polarity=0)

float64 v_f32_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, float64 income, bool64 predicate, bool polarity=0)

bfloat128 v_bf16_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, bfloat128 income, bool128 predicate, bool polarity=0)

int64 v_i32_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, int64 income, bool64 predicate, bool polarity=0)

uint64 v_u32_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, uint64 income, bool64 predicate, bool polarity=0)

short128 v_i16_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, short128 income, bool128 predicate, bool polarity=0)

ushort128 v_u16_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, ushort128 income, bool128 predicate, bool polarity=0)

char256 v_i8_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, char256 income, bool256 predicate, bool polarity=0)

uchar256 v_u8_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, uchar256 income, bool256 predicate, bool polarity=0)

bool256 v_i1_ld_tnsr_partial_vb(int5 ndx, int8_t tensor, int8_t size, int8_t offset, int switches, bool256 income, bool256 predicate, bool polarity=0)

Parameters

ndx	Tensor coordinates (SRC1).
tensor	Tensor number.
size	Size in elements minus 1.
offset	Offset in elements.
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

The loaded values.

Load tensor high

Represents LD_TNSR_HIGH instruction.

Intrinsics Prototype and Arguments

Intrinsics Definition

float64 v_f32_ld_tnsr_high_b(int5 ndx, const int8_t tensor, int switches=0, float64 income={}, bool predicate=1, bool polarity=0)

int64 v_i32_ld_tnsr_high_b(int5 ndx, const int8_t tensor, int switches=0, int64 income={}, bool predicate=1, bool polarity=0)

uint64 v_u32_ld_tnsr_high_b(int5 ndx, const int8_t tensor, int switches=0, uint64 income={}, bool predicate=1, bool polarity=0)

short128 v_i16_ld_tnsr_high_b(int5 ndx, const int8_t tensor, int switches=0, short128 income={}, bool predicate=1, bool polarity=0)

ushort128 v_u16_ld_tnsr_high_b(int5 ndx, const int8_t tensor, int switches=0, ushort128 income={}, bool predicate=1, bool polarity=0)

char256 v_i8_ld_tnsr_high_b(int5 ndx, const int8_t tensor, int switches=0, char256 income={}, bool predicate=1, bool polarity=0)

uchar256 v_u8_ld_tnsr_high_b(int5 ndx, const int8_t tensor, int switches=0, uchar256 income={}, bool predicate=1, bool polarity=0)

bool256 v_i1_ld_tnsr_high_b(int5 ndx, const int8_t tensor, int switches=0, bool256 income={}, bool predicate=1, bool polarity=0)

float64 v_f32_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, float64 income, bool64 predicate, bool polarity=0)

bfloat128 v_bf16_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, bfloat128 income, bool128 predicate, bool polarity=0)

int64 v_i32_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, int64 income, bool64 predicate, bool polarity=0)

uint64 v_u32_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, uint64 income, bool64 predicate, bool polarity=0)

short128 v_i16_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, short128 income, bool128 predicate, bool polarity=0)

ushort128 v_u16_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, ushort128 income, bool128 predicate, bool polarity=0)

char256 v_i8_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, char256 income, bool256 predicate, bool polarity=0)

uchar256 v_u8_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, uchar256 income, bool256 predicate, bool polarity=0)

bool256 v_i1_ld_tnsr_high_vb(int5 ndx, int8_t tensor, int switches, bool256 income, bool256 predicate, bool polarity=0)

bfloat128 v_bf16_ld_tnsr_high_b(int5 ndx, int8_t tensor, int switches=0, bfloat128 income={}, bool predicate=1, bool polarity=0)

Parameters

ndx	Tensor coordinates (SRC1).
tensor	Tensor number.
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

The loaded values.

Load tensor low

Represents LD_TNSR_LOW instruction.

Intrinsics Prototype and Arguments

Intrinsics Definition

float64 v_f32_ld_tnsr_low_b(int5 ndx, const int8_t tensor, int switches=0, float64 income={}, bool predicate=1, bool polarity=0)

int64 v_i32_ld_tnsr_low_b(int5 ndx, const int8_t tensor, int switches=0, int64 income={}, bool predicate=1, bool polarity=0)

uint64 v_u32_ld_tnsr_low_b(int5 ndx, const int8_t tensor, int switches=0, uint64 income={}, bool predicate=1, bool polarity=0)

short128 v_i16_ld_tnsr_low_b(int5 ndx, const int8_t tensor, int switches=0, short128 income={}, bool predicate=1, bool polarity=0)

ushort128 v_u16_ld_tnsr_low_b(int5 ndx, const int8_t tensor, int switches=0, ushort128 income={}, bool predicate=1, bool polarity=0)

char256 v_i8_ld_tnsr_low_b(int5 ndx, const int8_t tensor, int switches=0, char256 income={}, bool predicate=1, bool polarity=0)

uchar256 v_u8_ld_tnsr_low_b(int5 ndx, const int8_t tensor, int switches=0, uchar256 income={}, bool predicate=1, bool polarity=0)

bool256 v_i1_ld_tnsr_low_b(int5 ndx, const int8_t tensor, int switches=0, bool256 income={}, bool predicate=1, bool polarity=0)

float64 v_f32_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, float64 income, bool64 predicate, bool polarity=0)

bfloat128 v_bf16_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, bfloat128 income, bool128 predicate, bool polarity=0)

int64 v_i32_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, int64 income, bool64 predicate, bool polarity=0)

uint64 v_u32_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, uint64 income, bool64 predicate, bool polarity=0)

short128 v_i16_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, short128 income, bool128 predicate, bool polarity=0)

ushort128 v_u16_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, ushort128 income, bool128 predicate, bool polarity=0)

char256 v_i8_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, char256 income, bool256 predicate, bool polarity=0)

uchar256 v_u8_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, uchar256 income, bool256 predicate, bool polarity=0)

bool256 v_i1_ld_tnsr_low_vb(int5 ndx, int8_t tensor, int switches, bool256 income, bool256 predicate, bool polarity=0)

bfloat128 v_bf16_ld_tnsr_low_b(int5 ndx, int8_t tensor, int switches=0, bfloat128 income={}, bool predicate=1, bool polarity=0)

Parameters

ndx	Tensor coordinates (SRC1).
tensor	Tensor number.
switches	Instruction switches.
income	This value is returned if the predicate is false.
predicate	Predicate value for the instruction.
polarity	True if polarity of the predicate is inverted.

Returns

The loaded value.