lfp.s http://sakai.it.tdt.edu.vn/access/lessonbuilder/item/11364/group/44fd420...

## LSU EE 4720 -- Spring 2014 -- Computer Architecture

#
## MIPS Floating Point
#
# Time-stamp: <22 January 2014, 15:00:04 CST, [email protected]>

## Contents
#
# MIPS Floating-Point Instructions

## Objectives
#
# Floating Point
# Read and write MIPS programs using floating point instructions.

## Differences Between FP and Integer Instructions

#

# o Separate set of registers. (f0-f31)

# o Separate instructions to operate on those registers ..
# ... for arithmetic (add) ...
# ... and loads and stores.
# o Need for format conversion.

## Registers
#
## MIPS Registers
#
# MIPS has four sets of coprocessor registers.
# The integer (GPR) registers are NOT one of the four sets.
# Each set has 32 registers.
#
# Co-processor 0: Processor and system control.
# Co-processor 1: MIPS-32 floating-point
# Co-processor 2: Reserved for special-purpose designs.
# Co-processor 3: MIPS-64 floating-point
#

## Separate Floating Point Registers

#
# A feature of many RISC ISAs.
# Eases implementation.

################################################################################
## Floating Point Summary

## Representation of Floating-Point Numbers

#
# IEEE 754 Single- and Double-Precision FP Numbers
#
# Why IEEE 754?
# Like the metric system ..

1 of 5 21/03/2017 03:46 PM
lfp.s http://sakai.it.tdt.edu.vn/access/lessonbuilder/item/11364/group/44fd420...

# .. it's a good standard ..

# .. but much more importantly: many others use it.
#
# Data files containing FP numbers can be shared
# .. with other systems that implement IEEE 754 ..
# .. without conversion (except for byte order).

## Coprocessor 1 and The Floating-Point Registers

#
# Floating point handled by co-processor 1, one of 4 co-processors.
#
# MIPS floating point registers also called co-processor 1 registers.
# MIPS floating point instructions called co-processor 1 instructions.
#
# Registers named f0-f31.
# Each register is 32 bits. (For MIPS-32)

## Coprocessor 1 Instructions
#
# Coprocessor 1 instructions use coprocessor 1 (FP) registers.
# This includes instructions that do FP arithmetic ..
# .. and other types of instructions.
#
# Many coprocessor 1 instructions have "c1" in their names ..
# .. for example, lwc1.
#
# Coprocessor 1 *arithmetic* instructions *do not* have c1 in their names ..
# .. but they include a /completer/ that indicates data type ..
# .. for example, add.d, where ".d" is the completer.
#
# Completers:
#
# ".s" Single-Precision Floating Point (32 bits)
# ".d" Double-Precision Floating Point (64 bits)
# ".w" Integer (32 bits)

## Double-Precision Operands
#
# In MIPS-32 (including MIPS-I) FP registers are 32 bits.
# Instructions with ".d" operands get each operand from a pair of regs.
# Register numbers must be even.

# :Example:

add.d $f0, $f2, $f4 # {$f0,$f1} = { $f2, $f3 } + { $f4, $f5 }

add.d $f0, $f2, $f5 # ILLEGAL in MIPS 32, because f5 is odd.

## Immediate Operands and Constant Registers

#
# MIPS FP instructions do not take immediate values.
addi.s $f0, $f1, 2.3 # ILLEGAL, no immediate FP instructions.

# There is no counterpart of integer register 0 ($0 aka r0 aka $zero).

add.s $f0, $f1, $f2 # f0 has the sum,

2 of 5 21/03/2017 03:46 PM
lfp.s http://sakai.it.tdt.edu.vn/access/lessonbuilder/item/11364/group/44fd420...

## Types of Floating-Point Instructions

#
# Briefly here, in detail later.
#
#
## Arithmetic Operations
#
# Add double-precision (64-bit) operands.
#
# MIPS: add.d $f0, $f2, $f4

#
## Load and Store
#
# Load double (eight bytes into two consecutive registers).
#
# MIPS: ldc1 $f0, 8($t0)
#
#
## Move Between Register Files (E.g., integer to FP)
#
# MIPS: mtc1 $t0, $f0

## Format Conversion
#
# Convert from one format to another, e.g., integer to double.
#
# MIPS: cvt.d.w $f0, $f2
#
#
## Floating Point Condition Code Setting
#
# Compare and set condition code.
#
# MIPS: c.gt.d $f0, $f2
#
#
## Conditional Branch
#
# Branch on floating-point condition.
#
# MIPS: bc1f TARGET # Branch coprocessor 1 [condition code] false.

## FP Load and Store

# MIPS
#
# Load word in to coprocessor 1
lwc1 $f0, 4($t4) # $f0 = Mem[ $t4 + 4 ]

#
# Load double in to coprocessor 1
ldc1 $f0, 0($t4) # $f0 = Mem[ $t4 + 0 ]; $f1 = Mem[ $t4 + 4 ]
#
# Store word from coprocessor 1.
swc1 $f0, 4($t4) # $f0 = Mem[ $t4 + 4 ]
#

3 of 5 21/03/2017 03:46 PM
lfp.s http://sakai.it.tdt.edu.vn/access/lessonbuilder/item/11364/group/44fd420...

# Store double from coprocessor 1.

sdc1 $f0, 0($t4) # Mem[ $t4 + 0 ] = $f0; Mem[ $t4 + 4 ] = $f1

## Move Instructions

# MIPS
#
# Move to coprocessor 1
mtc1 $t0, $f0 # f0 = t0 Note that destination is *second* reg.
#
# Move from coprocessor 1.
mfc1 $t0, $f0

## Data Type Conversion

# Convert between floating-point and integer formats.

# NOTE: Values don't convert automatically, need to use these insn.

# MIPS
#
# To: s, d, w; From: s, d, w
#
# cvt.TO.FROM fd, fs
#
cvt.d.w $f0, $f2 # $f0 = convert_from_int_to_double( $f2 )

## Setting Condition Codes

# In preparation for a branch, set cond code based on FP comparison.

# MIPS
#
# Compare: fs COND ft
# COND: eq, gt, lt, le, ge
# FMT: s, d
#
# c.COND.FMT fs, ft
# Sets condition code to true or false.
# Condition is false if either operand is not a number.
#
c.lt.d $f0, $f2 # CC = $f0 < $f2
bc1t TARG # Branch if $f0 < $f2
nop
c.ge.d $f0, $f2 # CC = $f0 >= $f2
bc1t TARG2 # Branch if $f0 < $f2
nop
# Reachable?

## FP Branches

# MIPS
#
# Branch insn specifies whether CC register true or false.
#
bc1t TARG
nop
#

4 of 5 21/03/2017 03:46 PM
lfp.s http://sakai.it.tdt.edu.vn/access/lessonbuilder/item/11364/group/44fd420...

bc1f TARG
nop

5 of 5 21/03/2017 03:46 PM