Multiply-accumulate of bfloat16 datatypes

Matrix multiplications in which matrix A and B have bfloat16 data elements. More...

Overview

Matrix multiplications in which matrix A and B have bfloat16 data elements.

For an explanation how these operations works see Multiply Accumulate.

Multiplication of (4x8) with (8x4)
v16accfloat	mul_4x8_8x4 (v32bfloat16 a, v32bfloat16 b)
v16accfloat	negmul_4x8_8x4 (v32bfloat16 a, v32bfloat16 b)
v16accfloat	mac_4x8_8x4 (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1)
v16accfloat	msc_4x8_8x4 (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1)
v16accfloat	addmac_4x8_8x4 (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, v16accfloat acc2)
v16accfloat	addmsc_4x8_8x4 (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, v16accfloat acc2)

Multiplication of (4x8) with (8x4) with dynamic negation of multiplication result
v16accfloat	mul_4x8_8x4_conf (v32bfloat16 a, v32bfloat16 b, int sub_mul)
v16accfloat	negmul_4x8_8x4_conf (v32bfloat16 a, v32bfloat16 b, int sub_mul)

Multiplication of (4x8) with (8x4) with dynamic negation of multiplication result, zeroing of acc1, and negation of acc1
v16accfloat	mac_4x8_8x4_conf (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v16accfloat	msc_4x8_8x4_conf (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v16accfloat	addmac_4x8_8x4_conf (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, v16accfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v16accfloat	addmsc_4x8_8x4_conf (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, v16accfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)

Multiplication of (2x8) with (8x2) bfloat16 and complex bfloat16
v4caccfloat	mul_2x8_8x2 (v16bfloat16 a, v16cbfloat16 b)
v4caccfloat	negmul_2x8_8x2 (v16bfloat16 a, v16cbfloat16 b)
v4caccfloat	mac_2x8_8x2 (v16bfloat16 a, v16cbfloat16 b, v4caccfloat acc1)
v4caccfloat	msc_2x8_8x2 (v16bfloat16 a, v16cbfloat16 b, v4caccfloat acc1)
v4caccfloat	addmac_2x8_8x2 (v16bfloat16 a, v16cbfloat16 b, v4caccfloat acc1, v4caccfloat acc2)
v4caccfloat	addmsc_2x8_8x2 (v16bfloat16 a, v16cbfloat16 b, v4caccfloat acc1, v4caccfloat acc2)

Multiplication of (2x8) with (8x2) complex bfloat16 and bfloat16
v4caccfloat	mul_2x8_8x2 (v16cbfloat16 a, v16bfloat16 b)
v4caccfloat	negmul_2x8_8x2 (v16cbfloat16 a, v16bfloat16 b)
v4caccfloat	mac_2x8_8x2 (v16cbfloat16 a, v16bfloat16 b, v4caccfloat acc1)
v4caccfloat	msc_2x8_8x2 (v16cbfloat16 a, v16bfloat16 b, v4caccfloat acc1)
v4caccfloat	addmac_2x8_8x2 (v16cbfloat16 a, v16bfloat16 b, v4caccfloat acc1, v4caccfloat acc2)
v4caccfloat	addmsc_2x8_8x2 (v16cbfloat16 a, v16bfloat16 b, v4caccfloat acc1, v4caccfloat acc2)

Multiplication of (2x8) with (8x2) complex bfloat16 and complex bfloat16
v4caccfloat	mul_2x8_8x2 (v16cbfloat16 a, v16cbfloat16 b)
v4caccfloat	negmul_2x8_8x2 (v16cbfloat16 a, v16cbfloat16 b)
v4caccfloat	mac_2x8_8x2 (v16cbfloat16 a, v16cbfloat16 b, v4caccfloat acc1)
v4caccfloat	msc_2x8_8x2 (v16cbfloat16 a, v16cbfloat16 b, v4caccfloat acc1)
v4caccfloat	addmac_2x8_8x2 (v16cbfloat16 a, v16cbfloat16 b, v4caccfloat acc1, v4caccfloat acc2)
v4caccfloat	addmsc_2x8_8x2 (v16cbfloat16 a, v16cbfloat16 b, v4caccfloat acc1, v4caccfloat acc2)

Multiplication of (2x8) with (8x2) bfloat16 and complex bfloat16 with dynamic negation of multiplication result
v4caccfloat	mul_2x8_8x2_conf (v16bfloat16 a, v16cbfloat16 b, int sub_mul)
v4caccfloat	negmul_2x8_8x2_conf (v16bfloat16 a, v16cbfloat16 b, int sub_mul)

Multiplication of (2x8) with (8x2) complex bfloat16 and bfloat16 with dynamic negation of multiplication result
v4caccfloat	mul_2x8_8x2_conf (v16cbfloat16 a, v16bfloat16 b, int sub_mul)
v4caccfloat	negmul_2x8_8x2_conf (v16cbfloat16 a, v16bfloat16 b, int sub_mul)

Multiplication of (2x8) with (8x2) complex bfloat16 and complex bfloat16 with dynamic negation of multiplication result
v4caccfloat	mul_2x8_8x2_conf (v16cbfloat16 a, v16cbfloat16 b, int sub_mul)
v4caccfloat	negmul_2x8_8x2_conf (v16cbfloat16 a, v16cbfloat16 b, int sub_mul)

Multiplication of (2x8) with (8x2) bfloat16 and complex bfloat16 with dynamic negation of multiplication result, zeroing of acc1, and negation of acc1
v4caccfloat	mac_2x8_8x2_conf (v16bfloat16 v1, v16cbfloat16 v2, v4caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v4caccfloat	msc_2x8_8x2_conf (v16bfloat16 v1, v16cbfloat16 v2, v4caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v4caccfloat	addmac_2x8_8x2_conf (v16bfloat16 v1, v16cbfloat16 v2, v4caccfloat acc1, v4caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v4caccfloat	addmsc_2x8_8x2_conf (v16bfloat16 v1, v16cbfloat16 v2, v4caccfloat acc1, v4caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)

Multiplication of (2x8) with (8x2) complex bfloat16 and bfloat16 with dynamic negation of multiplication result, zeroing of acc1, and negation of acc1
v4caccfloat	mac_2x8_8x2_conf (v16cbfloat16 v1, v16bfloat16 v2, v4caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v4caccfloat	msc_2x8_8x2_conf (v16cbfloat16 v1, v16bfloat16 v2, v4caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v4caccfloat	addmac_2x8_8x2_conf (v16cbfloat16 v1, v16bfloat16 v2, v4caccfloat acc1, v4caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v4caccfloat	addmsc_2x8_8x2_conf (v16cbfloat16 v1, v16bfloat16 v2, v4caccfloat acc1, v4caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)

Multiplication of (2x8) with (8x2) complex bfloat16 and complex bfloat16 with dynamic negation of multiplication result, zeroing of acc1, and negation of acc1
v4caccfloat	mac_2x8_8x2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v4caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v4caccfloat	msc_2x8_8x2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v4caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v4caccfloat	addmac_2x8_8x2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v4caccfloat acc1, v4caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v4caccfloat	addmsc_2x8_8x2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v4caccfloat acc1, v4caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)

Channel by channel multiplication of (1x2) with (2x1)
v16accfloat	mul_elem_16_2 (v32bfloat16 a, v32bfloat16 b)
v16accfloat	negmul_elem_16_2 (v32bfloat16 a, v32bfloat16 b)
v16accfloat	mac_elem_16_2 (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1)
v16accfloat	msc_elem_16_2 (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1)
v16accfloat	addmac_elem_16_2 (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, v16accfloat acc2)
v16accfloat	addmsc_elem_16_2 (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, v16accfloat acc2)

Channel by channel multiplication of (1x2) with (2x1) with dynamic negation of multiplication result
v16accfloat	mul_elem_16_2_conf (v32bfloat16 a, v32bfloat16 b, int sub_mul)
v16accfloat	negmul_elem_16_2_conf (v32bfloat16 a, v32bfloat16 b, int sub_mul)

Channel by channel multiplication of (1x2) with (2x1) with dynamic negation of multiplication result, zeroing of acc1, and negation of acc1
v16accfloat	mac_elem_16_2_conf (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v16accfloat	msc_elem_16_2_conf (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v16accfloat	addmac_elem_16_2_conf (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, v16accfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v16accfloat	addmsc_elem_16_2_conf (v32bfloat16 a, v32bfloat16 b, v16accfloat acc1, v16accfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)

channel by channel multiplication of (1x2) with (2x1) complex bfloat16 and complex bfloat16
v8caccfloat	mul_elem_8_2 (v16cbfloat16 v1, v16cbfloat16 v2)
v8caccfloat	negmul_elem_8_2 (v16cbfloat16 v1, v16cbfloat16 v2)
v8caccfloat	mac_elem_8_2 (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1)
v8caccfloat	msc_elem_8_2 (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1)
v8caccfloat	addmac_elem_8_2 (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2)
v8caccfloat	addmsc_elem_8_2 (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2)

channel by channel multiplication of (1x2) with (2x1) bfloat16 and complex bfloat16
v8caccfloat	mul_elem_8_2 (v16bfloat16 v1, v16cbfloat16 v2)
v8caccfloat	negmul_elem_8_2 (v16bfloat16 v1, v16cbfloat16 v2)
v8caccfloat	mac_elem_8_2 (v16bfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1)
v8caccfloat	msc_elem_8_2 (v16bfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1)
v8caccfloat	addmac_elem_8_2 (v16bfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2)
v8caccfloat	addmsc_elem_8_2 (v16bfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2)

channel by channel multiplication of (1x2) with (2x1) complex bfloat16 and bfloat16
v8caccfloat	mul_elem_8_2 (v16cbfloat16 v1, v16bfloat16 v2)
v8caccfloat	negmul_elem_8_2 (v16cbfloat16 v1, v16bfloat16 v2)
v8caccfloat	mac_elem_8_2 (v16cbfloat16 v1, v16bfloat16 v2, v8caccfloat acc1)
v8caccfloat	msc_elem_8_2 (v16cbfloat16 v1, v16bfloat16 v2, v8caccfloat acc1)
v8caccfloat	addmac_elem_8_2 (v16cbfloat16 v1, v16bfloat16 v2, v8caccfloat acc1, v8caccfloat acc2)
v8caccfloat	addmsc_elem_8_2 (v16cbfloat16 v1, v16bfloat16 v2, v8caccfloat acc1, v8caccfloat acc2)

channel by channel multiplication of (1x2) with (2x1) for complex bfloat16 and complex bfloat16 with dynamic negation of multiplication result
v8caccfloat	mul_elem_8_2_conf (v16cbfloat16 v1, v16cbfloat16 v2, int sub_mul)
v8caccfloat	negmul_elem_8_2_conf (v16cbfloat16 v1, v16cbfloat16 v2, int sub_mul)

channel by channel multiplication of (1x2) with (2x1) for complex bfloat16 and complex bfloat16 with dynamic negation of multiplication result, zeroing of acc1, and negation of acc1
v8caccfloat	mac_elem_8_2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v8caccfloat	msc_elem_8_2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v8caccfloat	addmac_elem_8_2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v8caccfloat	submac_elem_8_2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v8caccfloat	addmsc_elem_8_2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v8caccfloat	submsc_elem_8_2_conf (v16cbfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)

channel by channel multiplication of (1x2) with (2x1) for complex cbfloat16 and bfloat16 with dynamic negation of multiplication result
v8caccfloat	mul_elem_8_2_conf (v16cbfloat16 v1, v16bfloat16 v2, int sub_mul)
v8caccfloat	negmul_elem_8_2_conf (v16cbfloat16 v1, v16bfloat16 v2, int sub_mul)

channel by channel multiplication of (1x2) with (2x1) for complex bfloat16 and bfloat16 with dynamic negation of multiplication result, zeroing of acc1, and negation of acc1
v8caccfloat	mac_elem_8_2_conf (v16cbfloat16 v1, v16bfloat16 v2, v8caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v8caccfloat	msc_elem_8_2_conf (v16cbfloat16 v1, v16bfloat16 v2, v8caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v8caccfloat	addmac_elem_8_2_conf (v16cbfloat16 v1, v16bfloat16 v2, v8caccfloat acc1, v8caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v8caccfloat	addmsc_elem_8_2_conf (v16cbfloat16 v1, v16bfloat16 v2, v8caccfloat acc1, v8caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)

channel by channel multiplication of (1x2) with (2x1) for bfloat16 and complex bfloat16 with dynamic negation of multiplication result
v8caccfloat	mul_elem_8_2_conf (v16bfloat16 v1, v16cbfloat16 v2, int sub_mul)
v8caccfloat	negmul_elem_8_2_conf (v16bfloat16 v1, v16cbfloat16 v2, int sub_mul)

channel by channel multiplication of (1x2) with (2x1) for bfloat16 and complex bfloat16 with dynamic negation of multiplication result, zeroing of acc1, and negation of acc1
v8caccfloat	mac_elem_8_2_conf (v16bfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v8caccfloat	msc_elem_8_2_conf (v16bfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, int zero_acc1, int sub_mul, int sub_acc1)
v8caccfloat	addmac_elem_8_2_conf (v16bfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)
v8caccfloat	addmsc_elem_8_2_conf (v16bfloat16 v1, v16cbfloat16 v2, v8caccfloat acc1, v8caccfloat acc2, int zero_acc1, int sub_mul, int sub_acc1, int sub_acc2)

Function Documentation

addmac_2x8_8x2() [1/3]

v4caccfloat addmac_2x8_8x2	(	v16bfloat16	a,
		v16cbfloat16	b,
		v4caccfloat	acc1,
		v4caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmac_2x8_8x2() [2/3]

v4caccfloat addmac_2x8_8x2	(	v16cbfloat16	a,
		v16bfloat16	b,
		v4caccfloat	acc1,
		v4caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmac_2x8_8x2() [3/3]

v4caccfloat addmac_2x8_8x2	(	v16cbfloat16	a,
		v16cbfloat16	b,
		v4caccfloat	acc1,
		v4caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmac_2x8_8x2_conf() [1/3]

v4caccfloat addmac_2x8_8x2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v4caccfloat	acc1,
		v4caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmac_2x8_8x2_conf() [2/3]

v4caccfloat addmac_2x8_8x2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v4caccfloat	acc1,
		v4caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmac_2x8_8x2_conf() [3/3]

v4caccfloat addmac_2x8_8x2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v4caccfloat	acc1,
		v4caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmac_4x8_8x4()

v16accfloat addmac_4x8_8x4	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		v16accfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmac_4x8_8x4_conf()

v16accfloat addmac_4x8_8x4_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		v16accfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmac_elem_16_2()

v16accfloat addmac_elem_16_2	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		v16accfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmac_elem_16_2_conf()

v16accfloat addmac_elem_16_2_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		v16accfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmac_elem_8_2() [1/3]

v8caccfloat addmac_elem_8_2	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmac_elem_8_2() [2/3]

v8caccfloat addmac_elem_8_2	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmac_elem_8_2() [3/3]

v8caccfloat addmac_elem_8_2	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmac_elem_8_2_conf() [1/3]

v8caccfloat addmac_elem_8_2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmac_elem_8_2_conf() [2/3]

v8caccfloat addmac_elem_8_2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmac_elem_8_2_conf() [3/3]

v8caccfloat addmac_elem_8_2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmsc_2x8_8x2() [1/3]

v4caccfloat addmsc_2x8_8x2	(	v16bfloat16	a,
		v16cbfloat16	b,
		v4caccfloat	acc1,
		v4caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmsc_2x8_8x2() [2/3]

v4caccfloat addmsc_2x8_8x2	(	v16cbfloat16	a,
		v16bfloat16	b,
		v4caccfloat	acc1,
		v4caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmsc_2x8_8x2() [3/3]

v4caccfloat addmsc_2x8_8x2	(	v16cbfloat16	a,
		v16cbfloat16	b,
		v4caccfloat	acc1,
		v4caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmsc_2x8_8x2_conf() [1/3]

v4caccfloat addmsc_2x8_8x2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v4caccfloat	acc1,
		v4caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmsc_2x8_8x2_conf() [2/3]

v4caccfloat addmsc_2x8_8x2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v4caccfloat	acc1,
		v4caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmsc_2x8_8x2_conf() [3/3]

v4caccfloat addmsc_2x8_8x2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v4caccfloat	acc1,
		v4caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmsc_4x8_8x4()

v16accfloat addmsc_4x8_8x4	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		v16accfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmsc_4x8_8x4_conf()

v16accfloat addmsc_4x8_8x4_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		v16accfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmsc_elem_16_2()

v16accfloat addmsc_elem_16_2	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		v16accfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmsc_elem_16_2_conf()

v16accfloat addmsc_elem_16_2_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		v16accfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmsc_elem_8_2() [1/3]

v8caccfloat addmsc_elem_8_2	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmsc_elem_8_2() [2/3]

v8caccfloat addmsc_elem_8_2	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmsc_elem_8_2() [3/3]

v8caccfloat addmsc_elem_8_2	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input

Returns

Result of operation

addmsc_elem_8_2_conf() [1/3]

v8caccfloat addmsc_elem_8_2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmsc_elem_8_2_conf() [2/3]

v8caccfloat addmsc_elem_8_2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

addmsc_elem_8_2_conf() [3/3]

v8caccfloat addmsc_elem_8_2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

mac_2x8_8x2() [1/3]

v4caccfloat mac_2x8_8x2	(	v16bfloat16	a,
		v16cbfloat16	b,
		v4caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

mac_2x8_8x2() [2/3]

v4caccfloat mac_2x8_8x2	(	v16cbfloat16	a,
		v16bfloat16	b,
		v4caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

mac_2x8_8x2() [3/3]

v4caccfloat mac_2x8_8x2	(	v16cbfloat16	a,
		v16cbfloat16	b,
		v4caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

mac_2x8_8x2_conf() [1/3]

v4caccfloat mac_2x8_8x2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v4caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

mac_2x8_8x2_conf() [2/3]

v4caccfloat mac_2x8_8x2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v4caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

mac_2x8_8x2_conf() [3/3]

v4caccfloat mac_2x8_8x2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v4caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

mac_4x8_8x4()

v16accfloat mac_4x8_8x4	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

mac_4x8_8x4_conf()

v16accfloat mac_4x8_8x4_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

mac_elem_16_2()

v16accfloat mac_elem_16_2	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

mac_elem_16_2_conf()

v16accfloat mac_elem_16_2_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

mac_elem_8_2() [1/3]

v8caccfloat mac_elem_8_2	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

mac_elem_8_2() [2/3]

v8caccfloat mac_elem_8_2	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v8caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

mac_elem_8_2() [3/3]

v8caccfloat mac_elem_8_2	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

mac_elem_8_2_conf() [1/3]

v8caccfloat mac_elem_8_2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result

Returns

Result of operation

mac_elem_8_2_conf() [2/3]

v8caccfloat mac_elem_8_2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v8caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result

Returns

Result of operation

mac_elem_8_2_conf() [3/3]

v8caccfloat mac_elem_8_2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result

Returns

Result of operation

msc_2x8_8x2() [1/3]

v4caccfloat msc_2x8_8x2	(	v16bfloat16	a,
		v16cbfloat16	b,
		v4caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

msc_2x8_8x2() [2/3]

v4caccfloat msc_2x8_8x2	(	v16cbfloat16	a,
		v16bfloat16	b,
		v4caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

msc_2x8_8x2() [3/3]

v4caccfloat msc_2x8_8x2	(	v16cbfloat16	a,
		v16cbfloat16	b,
		v4caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

msc_2x8_8x2_conf() [1/3]

v4caccfloat msc_2x8_8x2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v4caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

msc_2x8_8x2_conf() [2/3]

v4caccfloat msc_2x8_8x2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v4caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

msc_2x8_8x2_conf() [3/3]

v4caccfloat msc_2x8_8x2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v4caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

msc_4x8_8x4()

v16accfloat msc_4x8_8x4	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

msc_4x8_8x4_conf()

v16accfloat msc_4x8_8x4_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

msc_elem_16_2()

v16accfloat msc_elem_16_2	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

msc_elem_16_2_conf()

v16accfloat msc_elem_16_2_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		v16accfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1

Returns

Result of operation

msc_elem_8_2() [1/3]

v8caccfloat msc_elem_8_2	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

msc_elem_8_2() [2/3]

v8caccfloat msc_elem_8_2	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v8caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

msc_elem_8_2() [3/3]

v8caccfloat msc_elem_8_2	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input

Returns

Result of operation

msc_elem_8_2_conf() [1/3]

v8caccfloat msc_elem_8_2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result

Returns

Result of operation

msc_elem_8_2_conf() [2/3]

v8caccfloat msc_elem_8_2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		v8caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result

Returns

Result of operation

msc_elem_8_2_conf() [3/3]

v8caccfloat msc_elem_8_2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result

Returns

Result of operation

mul_2x8_8x2() [1/3]

v4caccfloat mul_2x8_8x2	(	v16bfloat16	a,
		v16cbfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

mul_2x8_8x2() [2/3]

v4caccfloat mul_2x8_8x2	(	v16cbfloat16	a,
		v16bfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

mul_2x8_8x2() [3/3]

v4caccfloat mul_2x8_8x2	(	v16cbfloat16	a,
		v16cbfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

mul_2x8_8x2_conf() [1/3]

v4caccfloat mul_2x8_8x2_conf	(	v16bfloat16	a,
		v16cbfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result

Returns

Result of operation

mul_2x8_8x2_conf() [2/3]

v4caccfloat mul_2x8_8x2_conf	(	v16cbfloat16	a,
		v16bfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result

Returns

Result of operation

mul_2x8_8x2_conf() [3/3]

v4caccfloat mul_2x8_8x2_conf	(	v16cbfloat16	a,
		v16cbfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result

Returns

Result of operation

mul_4x8_8x4()

v16accfloat mul_4x8_8x4	(	v32bfloat16	a,
		v32bfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

mul_4x8_8x4_conf()

v16accfloat mul_4x8_8x4_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result

Returns

Result of operation

mul_elem_16_2()

v16accfloat mul_elem_16_2	(	v32bfloat16	a,
		v32bfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

mul_elem_16_2_conf()

v16accfloat mul_elem_16_2_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result

Returns

Result of operation

mul_elem_8_2() [1/3]

v8caccfloat mul_elem_8_2	(	v16bfloat16	v1,
		v16cbfloat16	v2
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

mul_elem_8_2() [2/3]

v8caccfloat mul_elem_8_2	(	v16cbfloat16	v1,
		v16bfloat16	v2
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

mul_elem_8_2() [3/3]

v8caccfloat mul_elem_8_2	(	v16cbfloat16	v1,
		v16cbfloat16	v2
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

mul_elem_8_2_conf() [1/3]

v8caccfloat mul_elem_8_2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask of multiplication result

Returns

Result of operation

mul_elem_8_2_conf() [2/3]

v8caccfloat mul_elem_8_2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask of multiplication result

Returns

Result of operation

mul_elem_8_2_conf() [3/3]

v8caccfloat mul_elem_8_2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask of multiplication result

Returns

Result of operation

negmul_2x8_8x2() [1/3]

v4caccfloat negmul_2x8_8x2	(	v16bfloat16	a,
		v16cbfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

negmul_2x8_8x2() [2/3]

v4caccfloat negmul_2x8_8x2	(	v16cbfloat16	a,
		v16bfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

negmul_2x8_8x2() [3/3]

v4caccfloat negmul_2x8_8x2	(	v16cbfloat16	a,
		v16cbfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

negmul_2x8_8x2_conf() [1/3]

v4caccfloat negmul_2x8_8x2_conf	(	v16bfloat16	a,
		v16cbfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result

Returns

Result of operation

negmul_2x8_8x2_conf() [2/3]

v4caccfloat negmul_2x8_8x2_conf	(	v16cbfloat16	a,
		v16bfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result

Returns

Result of operation

negmul_2x8_8x2_conf() [3/3]

v4caccfloat negmul_2x8_8x2_conf	(	v16cbfloat16	a,
		v16cbfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result

Returns

Result of operation

negmul_4x8_8x4()

v16accfloat negmul_4x8_8x4	(	v32bfloat16	a,
		v32bfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

negmul_4x8_8x4_conf()

v16accfloat negmul_4x8_8x4_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result. If a bit of sub_mul is set the corresponding vector lane of the output accumulator will be negated.

Returns

Result of operation

negmul_elem_16_2()

v16accfloat negmul_elem_16_2	(	v32bfloat16	a,
		v32bfloat16	b
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

negmul_elem_16_2_conf()

v16accfloat negmul_elem_16_2_conf	(	v32bfloat16	a,
		v32bfloat16	b,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask for multiplication result. If a bit of sub_mul is set the corresponding vector lane of the output accumulator will be negated.

Returns

Result of operation

negmul_elem_8_2() [1/3]

v8caccfloat negmul_elem_8_2	(	v16bfloat16	v1,
		v16cbfloat16	v2
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

negmul_elem_8_2() [2/3]

v8caccfloat negmul_elem_8_2	(	v16cbfloat16	v1,
		v16bfloat16	v2
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

negmul_elem_8_2() [3/3]

v8caccfloat negmul_elem_8_2	(	v16cbfloat16	v1,
		v16cbfloat16	v2
	)

Parameters

a	Matrix A
b	Matrix B

Returns

Result of operation

negmul_elem_8_2_conf() [1/3]

v8caccfloat negmul_elem_8_2_conf	(	v16bfloat16	v1,
		v16cbfloat16	v2,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask of multiplication result

Returns

Result of operation

negmul_elem_8_2_conf() [2/3]

v8caccfloat negmul_elem_8_2_conf	(	v16cbfloat16	v1,
		v16bfloat16	v2,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask of multiplication result

Returns

Result of operation

negmul_elem_8_2_conf() [3/3]

v8caccfloat negmul_elem_8_2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		int	sub_mul
	)

Parameters

a	Matrix A
b	Matrix B
sub_mul	Negation mask of multiplication result

Returns

Result of operation

submac_elem_8_2_conf()

v8caccfloat submac_elem_8_2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation

submsc_elem_8_2_conf()

v8caccfloat submsc_elem_8_2_conf	(	v16cbfloat16	v1,
		v16cbfloat16	v2,
		v8caccfloat	acc1,
		v8caccfloat	acc2,
		int	zero_acc1,
		int	sub_mul,
		int	sub_acc1,
		int	sub_acc2
	)

Parameters

a	Matrix A
b	Matrix B
acc1	Accumulator 1 input
acc2	Accumulator 2 input
zero_acc1	Zeroing mask for acc1
sub_mul	Negation mask of multiplication result
sub_acc1	Negation mask of acc1
sub_acc2	Negation mask of acc2

Returns

Result of operation