FLT.H
Float Less Than (Half)
FLT.H rd, rs1, rs2
Sets integer rd to 1 if half-precision rs1 is less than rs2, else 0.
Details
Performs a half-precision (16-bit) floating-point less-than comparison and writes 1 (true) or 0 (false) to integer rd. NaN inputs produce 0 (unordered), except FEQ which raises invalid-operation if either operand is a signalling NaN.
Pseudocode Operation
R[rd] = (F[rs1] < F[rs2]) ? 1 : 0;
Example
FLT.H t0, a0, a1
Encoding
Binary Layout
1010000
31:25
rs2
24:20
rs1
19:15
001
14:12
rd
11:7
1010011
6:0
Operands
-
rd
Dest (Int) -
rs1
Src 1 (Half) -
rs2
Src 2 (Half)