x86doc › KNOTW/KNOTB/KNOTQ/KNOTD - NOT Mask Register

KNOTW/KNOTB/KNOTQ/KNOTD—NOT Mask Register

Opcode/Instruction	Op/En	64/32 bit Mode Support	CPUID Feature Flag	Description
VEX.L0.0F.W0 44 /r KNOTW k1, k2	RR	V/V	AVX512F	Bitwise NOT of 16 bits mask k2.
VEX.L0.66.0F.W0 44 /r KNOTB k1, k2	RR	V/V	AVX512DQ	Bitwise NOT of 8 bits mask k2.
VEX.L0.0F.W1 44 /r KNOTQ k1, k2	RR	V/V	AVX512BW	Bitwise NOT of 64 bits mask k2.
VEX.L0.66.0F.W1 44 /r KNOTD k1, k2	RR	V/V	AVX512BW	Bitwise NOT of 32 bits mask k2.

Instruction Operand Encoding

Op/En	Operand 1	Operand 2
RR	ModRM:reg (w)	ModRM:r/m (r, ModRM:[7:6] must be 11b)

Description

Performs a bitwise NOT of vector mask k2 and writes the result into vector mask k1.

Operation

KNOTW

DEST[15:0] := BITWISE NOT SRC[15:0]
DEST[MAX_KL-1:16] := 0

KNOTB

DEST[7:0] := BITWISE NOT SRC[7:0]
DEST[MAX_KL-1:8] := 0

KNOTQ

DEST[63:0] := BITWISE NOT SRC[63:0]
DEST[MAX_KL-1:64] := 0

KNOTD

DEST[31:0] := BITWISE NOT SRC[31:0]
DEST[MAX_KL-1:32] := 0

Intel C/C++ Compiler Intrinsic Equivalent

KNOTW __mmask16 _mm512_knot(__mmask16 a);

Flags Affected

None.

SIMD Floating-Point Exceptions

None.

Other Exceptions

See Table 2-63, “TYPE K20 Exception Definition (VEX-Encoded OpMask Instructions w/o Memory Arg).”