[编译原理-鲸书]AdvancedCompilerDesignandImplementation - CSDN文库

2星需积分: 13 133 浏览量更新于2023-05-24 评论收藏 63.77MB PDF 举报

身份认证购VIP最低享 7 折!

领优惠券(最高得80元）

资源详情

资源评论

资源推荐

\"\"

'

,)

-)

.) .)

.)

..)

-)

, ,)

. ,II j)))

,)

D)

In-line

expansion

Leaf-routine

optimization

Shrink

wrapping

Machine idioms

Tail

merging

Branch

optimizations

and conditional

moves

Dead-code

elimination

Software

pipelining,

with

loop

unrolling,

variable

expansion, register

renaming,

and hierarchical reduction

Basic-blockand

branch

scheduling

1

Register

allocation

by graph coloring

Basic-block and branch

scheduling

2

Intraprocedural

I-cache

optimization

Instruction

prefetching

Data

prefetching

Branch

prediction)

,)

Interprocedural register

allocation

E

Aggregation

of

global

references

Interprocedural

I-cache

optimization)

(to

constant

folding,

algebraic

simplifications,

and

reassociation)

A)

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

- - - - - - - - - - - - - - - - I)

B,

C These

optimizations

are

typically performed

on medium-

or

low-level intermediate

code,

depending

on the

overall

organization

of

the

compiler.

If

code

selectionis done before all

optimizations

other than those

in

box

A

(known

as the \"low-level\" model of

optimizer

struc-

ture),

then these

optimizations

are

performed

on low-levelcode.

If,

on the other

hand,

some

optimizations

are

performed

on a

medium-level,

relatively

machine-independent

intermedi-

ate

code

and others are

performed

on low-level code after code

generation

(known

as

the

\"mixed\"

model),

then

these

optimizations

are

generally

done on

the medium-level interme-

diate code.

The

branches from Cl to

C2

and C3

represent

a

choiceof

the

method

used

to

perform

essentially

the same

optimization (namely, moving computations

to

places

where

they

are

per-

formed less

frequently

without

changing

the

semantics of

the

program).

They

also

represent

a choiceof

the

data-flow

analyses

used to

perform

the

optimization.

D

These

optimizations

are almost

always

done on

a low-level form

of code-one that

may

be

quite machine-dependent (e.g.,

a structured

assemblylanguage)

or that

may

be

somewhat

more

general,

such

as

the

low-level intermediate code

used

in

this

book-because

they

require

that

addresses

have

been turned

into

the

form

required by

the

target processor

and because

severalof them

require

low-level

control-flow code.

E

These

optimizations

are

performed

at

link

time,

so

they operate

on relocatable

object

code.

Three

optimizations, namely,

constant

folding, algebraic simplification,

and

reassociation,

are

in

boxes connected to the other

phases

of the

optimization process by

dotted

lines

because

they

are

best structured

as

subroutines

that can

beinvoked whenever

they

are needed.

A

version

of

this

diagram appears

in

Chapters

1

and

11

through

20 to

guide

the

reader

in

ordering optimizer components

in

a

compiler.)))

剩余886页未读，继续阅读

评论1

weixin_43515735

2019-02-17

还行，就是没目录

wencai_chang

粉丝: 1
资源: 12

会员权益专享

图片转文字

全年可省5，000元立即开通

最新资源

资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！点击此处反馈