cmpf.d | CS+ V6.01.00

cmpf.d

Floating-point compare (double)

[Syntax]

-	cmpf.d imm4, reg1, reg2, cc#3

-	cmpfcnd.d reg1, reg2

The following can be specified for imm4:

-	Absolute expression having a value up to 4 bits

[Function]

-	Syntax "cmpf.d imm4, reg1, reg2, cc#3"

The content in double-precision floating-point format in the register pair specified by reg2 is compared with the content in double-precision floating-point format in the register pair specified by reg1, via the imm4 comparison condition. The result (1 if true; 0 if false) is set in the condition bit (CC(7:0) bits; bits 31-24) in the FPSR register specified via cc#3. If cc#3 is omitted, it is set in the CC0 bit (bit 24).

-	Syntax "cmpfcnd.d reg1, reg2"

Via cmpfcnd.d, a corresponding "cmpf.d" instruction is generated (see "Table 5.50 cmpfcnd.d Instruction List" for details), and expanded in the format "cmpf.d imm4, reg1, reg2, cc#3". The content in single-precision floating-point format in the register pair specified by reg2 is compared with the content in single-precision floating-point format in the register pair specified by reg1, via the comparison condition. The result (1 if true; 0 if false) is set in the condition bit (CC(7:0) bits; bits 31-24) in the FPSR register specified via cc#3. If cc#3 is omitted, it is set in the CC0 bit (bit 24).

[Description]

-	If the instruction is executed in syntax "cmpf.d imm4, reg1, reg2, cc#3", the assembler generates one cmpf.d machine instruction.

-	If the instruction is executed in syntax "cmpfcnd.d reg1, reg2", the assembler generates the corresponding cmpf.d instruction (see "Table 5.50 cmpfcnd.d Instruction List") and expands it to syntax "cmpf.d imm4, reg1, reg2, cc#3".

Table 5.50

cmpfcnd.d Instruction List

Instruction	Condition	Meaning of Condition	Instruction Expansion
cmpff.d	FALSE	Always false	cmpf.d 0x0
cmpfun.d	Unordered	At least one of reg1 and reg2 is a non-number	cmpf.d 0x1
cmpfeq.d	reg2 = reg1	Neither is a non-number, and they are equal	cmpf.d 0x2
cmpfueq.d	reg2 ?= reg1	At least one is a non-number, or they are equal	cmpf.d 0x3
cmpfolt.d	reg2 < reg1	Neither is a non-number, and less than	cmpf.d 0x4
cmpfult.d	reg2 ?< reg1	At least one is a non-number, or less than	cmpf.d 0x5
cmpfole.d	reg2 <= reg1	Neither is a non-number, and less than or equal	cmpf.d 0x6
cmpfule.d	reg2 ?<= reg1	At least one is a non-number, or less than or equal	cmpf.d 0x7
cmpfsf.d	FALSE	Always false	cmpf.d 0x8
cmpfngle.d	Unordered	At least one of reg1 and reg2 is a non-number	cmpf.d 0x9
cmpfseq.d	reg2 = reg1	Neither is a non-number, and they are equal	cmpf.d 0xA
cmpfngl.d	reg2 ?= reg1	At least one is a non-number, or they are equal	cmpf.d 0xB
cmpflt.d	reg2 < reg1	Neither is a non-number, and less than	cmpf.d 0xC
cmpfnge.d	reg2 ?< reg1	At least one is a non-number, or less than	cmpf.d 0xD
cmpfle.d	reg2 <= reg1	Neither is a non-number, and less than or equal	cmpf.d 0xE
cmpfngt.d	reg2 ?<= reg1	At least one is a non-number, or less than or equal	cmpf.d 0xF

Remark

?: Unordered

[Caution]

-	If an absolute expression having a value exceeding 4 bits is specified as imm4 of the cmpf.d instruction, the following message is output, and assembly continues using the lower 4 bits of the specified value.

W0550011 : illegal operand (range error in immediate).