VFCLASS

Operation

Writes a floating-point classification bitmap for each active element.

Assembler Syntax

VFCLASS.{H|S|D}(z) Pn(p), Vn(v)

VFCLASS.{H|S|D}(z) Pn(p), <ea>(e), Vn(v)

VFCLASS.{H|S|D}(z) Pn(p), Vn(v), <ea>(e)

Privilege

Unprivileged

Required CPUID flags

FP

VECTOR

VECTORFP

For each active H, S, or D element, bits 0 through 9 of the result form a one-hot bitmap for negative infinity, negative normal, negative subnormal, negative zero, positive zero, positive subnormal, positive normal, positive infinity, signaling NaN, and quiet NaN, respectively. All higher bits of the destination element are zero.

The register form classifies each old destination lane and writes the result back to that lane. The memory-source form classifies each active memory element and writes the result to the corresponding destination lane. An inactive destination lane retains its previous value.

The memory-destination form classifies each active source-register lane and writes the result to the corresponding memory element. An inactive memory element retains its previous value, making the destination transaction a lane-wise read-modify-write.

2.75in

Encodings

VFCLASS.{H|S|D}(z) Pn(p), Vn(v)

Format: Instruction format for VFCLASS.{H|S|D}(z) Pn(p), Vn(v)

size — allowed: 0x1..0x3 [reserved_vector_element_type]

2.75in VFCLASS.{H|S|D}(z) Pn(p), <ea>(e), Vn(v)

Format: Instruction format for VFCLASS.{H|S|D}(z) Pn(p), <ea>(e), Vn(v)

size — allowed: 0x1..0x3 [reserved_vector_element_type]

2.75in VFCLASS.{H|S|D}(z) Pn(p), Vn(v), <ea>(e)

Format: Instruction format for VFCLASS.{H|S|D}(z) Pn(p), Vn(v), <ea>(e)

size — allowed: 0x1..0x3 [reserved_vector_element_type]

dst — excluded: immediate [invalid_destination]