User’s Manual

IBM PowerPC 750GX and 750GL RISC Microprocessor

For software compatibility, the other two mode encodings, imprecise-nonrecoverable mode and imprecise -recoverable mode, default to the precise mode.

Note: For the 750GX, the ignore-exceptions mode allows floating-point instructions to complete earlier and, thus, might provide better performance than the precise-exception mode.

•The floating-point load-and-store indexed instructions (lfsx, lfsux, lfdx, lfdux, stfsx, stfsux, stfdx, stfdux) are invalid when the Rc bit is one. In the 750GX, executing one of these invalid instruction forms causes CR0 to be set to an undefined value.

Floating-Point Load Instructions

There are two forms of the floating-point load instruction—single-precision and double-precision. The behavior of double-precision floating-point load instructions, and the behavior of single-precision floating- point load instructions are described here. Single-precision floating-point load instructions convert single- precision data to double-precision format before loading an operand into an FPR.

The PowerPC Architecture defines a load with update instruction with rA = 0 as an invalid form.

Table 2-23summarizes the single-precision and double-precision floating-point load instructions.

Table 2-23. Floating-Point Load Instructions

Name

Mnemonic

Syntax

 

 

 

 

 

 

Load Floating-Point Single

lfs

frD,d(rA)

 

 

 

Load Floating-Point Single Indexed

lfsx

frD,rA,rB

 

 

 

Load Floating-Point Single with Update

lfsu

frD,d(rA)

 

 

 

Load Floating-Point Single with Update Indexed

lfsux

frD,rA,rB

 

 

 

Load Floating-Point Double

lfd

frD,d(rA)

 

 

 

Load Floating-Point Double Indexed

lfdx

frD,rA,rB

 

 

 

Load Floating-Point Double with Update

lfdu

frD,d(rA)

 

 

 

Load Floating-Point Double with Update Indexed

lfdux

frD,rA,rB

 

 

 

Floating-Point Store Instructions

This section describes floating-point store instructions. There are three basic forms of the store instruction— single-precision, double-precision, and integer. The integer form is supported by the optional stfiwx instruc- tion. The behavior of double-precision floating-point store instructions, and the behavior of single-precision floating-point store instructions are described here. Single-precision floating-point store instructions convert double-precision data to single-precision format before storing the operands.

Programming Note: After power-on-reset, never store data from the Floating Point Register file because the file contains unset data and might have invalid formatted floating-point data. Always initialize the Floating Point Register file with valid floating-point data before continuing after a power-on-reset,.

Programming Model

gx_02.fm.(1.2)

Page 104 of 377

March 27, 2006