Floating-Point Multiply-Add Instructions; Floating-Point Rounding And Conversion Instructions - IBM PowerPC 604 User Manual

Table 2·14. Floatlng-Polnt Arithmetic Instructions {Continued)
Name Mnemonic Operand Syntax
Floating Divide Single ldlvs (ldlvs.)
lrD,lrA,lrB
Floating Square Roal (Double-Precision) lsqn (lsqn.)
frD,lrB
Floating Square Roal Single lsqns (lsqns.)
lrD,lrB
Floating
Re~rocal
Estimate Single
Ires (Ires.)
lrD,lrB
Floating
Re~rocal
Square Roal Estimate lrsqne (lrsqne.)
frD,lrB
Floating Select fsel lrD,lrA,lrC,frB
All single-precision arithmetic instructions are performed using a double-precision format
The floating-point architecture is a single-pass implementation for double-precision
products. In most cases, a single-precision instruction using only single-precision
operands, in double-precision format, has the same latency as its double-precision
equivalent.
2.3.4.2.2 Floating-Point Multiply-Add Instructions
These instructions combine multiply and add operations without an intermediate rounding
operation. The floating-point multiply-add instructions are summarized in Table 2-15.
Table 2·15. Floatlng-Polnt Multlply·Add Instructions
Name Mnemonic Operand Syntax
Floating Multiply-Add (Double-Precision) fmadd (lmadd.) frD,frA,lrC,frB
Floating Multiply-Add Single
fmadds (lmadds.)
frD,frA,frC,frB
Floating Multfply-Subtract (Double-Precision) lmsub (lmsub.) frD,frA,frC,frB
Floating Multfply-Subtract Single
fmsubs (fmsubs.)
frD,lrA,frC,lrB
Floating Negative Multfply-Add (Double-Precision) fnmadd (fnmadd.) lrD,frA,frC,lrB
Floating Negative Multfply-Add Single lnmadds (fnmadds.) lrD,lrA,frC,frB
Floating Negative Multfply-Subtrad (Double-Precision) lnmsub (fnmsub.) lrD,frA,frC,frB
Floating Negative
Mult~ly-Subtrad
Single
fnmsubs (fnmsubs.)
frD,lrA,lrC,frB
2.3.4.2.3 Floating-Point Rounding and Conversion Instructions
The Floating Round to Single-Precision (frsp) instruction is used to truncate a 64-bit
double-precision number to a 32-bit single-precision floating-point number. The
floating-point convert instructions convert a 64-bit double-precision floating-point number
to a 32-bit signed integer number.
Chapter 2- PowerPC 604 Pr008880r Programming Model
2-31
-