4.8 FPU Instruction Encoding and Opcode Field Description
Field | Description |
---|---|
lit16 | 16-bit unsigned literal Î {0...65535} |
index | Literal address index into 32 entry SP and DP constant tables |
.s | 32-bit Single Precision selection |
.d | 64-bit Double Precision selection |
Fd | One of 32 destination floating-point registers Î {F0..F31} (Register Direct) |
Fb | One of 32 source floating-point registers Î {F0..F31} (Register Direct) |
Fs | One of 32 source floating-point registers Î {F0..F31} (Register Direct) |
SUB, INF, FN, FZ, FNAN GT, LT, EQ, UN SUBO, HUGI, INX, UDF, OVF, DIV0, INVAL |
FPU status bits (FSR) Note: Flags for exceptions that cannot be
signaled by an instruction are always cleared by that
instruction (sticky flags are unaffected)
|
Field | Description |
---|---|
P | Selects Single Precision (P=0) or Double Precision (P=1) operation |
R | Selects between FPU coprocessor special registers (R=1) or F-regs (R=0) |
U | Unused (don’t care) instruction bit. Assembler to assign ‘0’ |
ddddd |
R=0: FPU coprocessor Fd destination register select: 00000=F0; 11111=F31 R=1: FPU coprocessor special register select |
eee | FPU rounding mode selection |
kkkk kkkk kkkk kkkk | 16-bit literal field, constant data |
sssss |
R=0: FPU coprocessor Fs source register select: 00000=F0; 11111=F31 R=1: FPU coprocessor special register select |