godzil/ack
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

changed font 5 references to font CW references

Browse Source
This commit is contained in:
ceriel
1988-04-18 13:34:29 +00:00
parent a9ad34c8db
commit 71efb88a36
8 changed files with 67 additions and 67 deletions
Download Patch File Download Diff File Expand all files Collapse all files

64
doc/top/top.n
View File

@@ -82,7 +82,7 @@ machine-dependent description table (see figure 1.).
So the major part of the code of a target optimizer is
shared among all target optimizers.
.DS
.ft 5
.ft CW
|-------------------------|
@@ -164,7 +164,7 @@ ANY matches every instruction mnemonic.
.nf
Examples of mnemonic descriptions:
.ft 5
.ft CW
add
sub.l
@@ -177,7 +177,7 @@ An operand can also be described by a string constant.
.nf
Examples:
.ft 5
.ft CW
(sp)+
r5
@@ -192,7 +192,7 @@ Each such declaration defines the name of a variable and
a \fIrestriction\fR to which its value is subjected.
.nf
Example of variable declarations:
.ft 5
.ft CW
CONST { VAL[0] == '$' };
REG { VAL[0] == 'r' && VAL[1] >= '0' && VAL[1] <= '3' &&
@@ -206,7 +206,7 @@ a null-terminated string.
An operand description given via a variable name matches an
actual operand if the actual operand obeys the associated restriction.
.nf
.ft 5
.ft CW
CONST matches $1, $-5, $foo etc.
REG matches r0, r1, r2 and r3
@@ -224,7 +224,7 @@ These procedures must be added to the table after the patterns.
.nf
Example:
.ft 5
.ft CW
FERMAT_NUMBER { VAL[0] == '$' && is_fermat_number(&VAL[1]) };
@@ -238,7 +238,7 @@ The most general form allowed is:
string_constant1 variable_name string_constant2
Example:
.ft 5
.ft CW
(REG)+ matches (r0)+, (r1)+, (r2)+ and (r3)+
@@ -268,19 +268,19 @@ the optional constraint C is satisfied, i.e. it evaluates to TRUE.
.LP
.nf
The pattern:
.ft 5
.ft CW
dec REG : move.b CONST,(REG)
.ft R
matches:
.ft 5
.ft CW
dec r0 : move.b $4,(r0)
.ft R
but not:
.ft 5
.ft CW
dec r0 : move.b $4,(r1)
@@ -292,7 +292,7 @@ extra names for a register should be declared, all sharing
the same restriction.
.nf
Example:
.ft 5
.ft CW
REG1,REG2 { VAL[0] == 'r' && ..... };
@@ -305,13 +305,13 @@ the parameter restrictions).
The expression may refer to the variables and to ANY.
.nf
Example:
.ft 5
.ft CW
move REG1,REG2 { REG1[1] == REG2[1] + 1 }
.ft R
matches
.ft 5
.ft CW
move r1,r0
move r2,r1
@@ -338,7 +338,7 @@ Vax examples
Suppose the table contains the following declarations:
.nf
.ft 5
.ft CW
X, LOG { TRUE };
LAB { VAL[0] == 'L' }; /* e.g. L0017 */
A { no_side_effects(VAL) };
@@ -353,7 +353,7 @@ These procedures must be supplied by the table-writer and must be
included in the table.
.PP
.nf
.ft 5
.ft CW
\fIentry:\fP addl3 X,A,A -> addl2 X,A;
.ft R
@@ -362,7 +362,7 @@ This entry changes a 3-operand instruction into a cheaper 2-operand
instruction.
An optimization like:
.nf
.ft 5
.ft CW
addl3 r0,(r2)+,(r2)+ -> addl2 r0,(r2)+
@@ -373,7 +373,7 @@ Hence the second argument is required to
be side-effect free.
.PP
.nf
.ft 5
.ft CW
\fIentry:\fP addw2 $-NUM,X -> subw2 $NUM,X;
.ft R
@@ -384,7 +384,7 @@ because constants in the range 0 to 63 are represented
very efficiently on the Vax.
.PP
.nf
.ft 5
.ft CW
\fIentry:\fP bitw $NUM,A : jneq LAB
{ is_poweroftwo(NUM,LOG) } -> jbs $LOG,A,LAB;
@@ -395,7 +395,7 @@ x and y.
A "jbs n,x,l" branches to l if bit n of x is set.
So, for example, the following transformation is possible:
.nf
.ft 5
.ft CW
bitw $32,r0 : jneq L0017 -> jbs $5,r0,L0017
@@ -413,7 +413,7 @@ PDP-11 examples
Suppose we have the following declarations:
.nf
.ft 5
.ft CW
X { TRUE };
A { no_side_effects(VAL) };
L1, L2 { VAL[0] == 'I' };
@@ -426,7 +426,7 @@ The implementation of "no_side_effects" may of course
differ for the PDP-11 and the Vax.
.PP
.nf
.ft 5
.ft CW
\fIentry:\fP mov REG,A : ANY A,X -> mov REG,A : ANY REG,X ;
.ft R
@@ -436,7 +436,7 @@ If A and REG hold the same value (which is true after "mov REG,A")
and A is used as source (first) operand, it is cheaper to use REG instead.
.PP
.nf
.ft 5
.ft CW
\fIentry:\fP jeq L1 : jbr L2 : labdef L1 -> jne L2 : labdef L1;
.ft R
@@ -447,7 +447,7 @@ As the target optimizer has to know how such a definition
looks like, this must be expressed in the table (see Appendix A).
.PP
.nf
.ft 5
.ft CW
\fIentry:\fP add $01,X { carry_dead(REST) } -> inc X;
.ft R
@@ -487,27 +487,27 @@ backwards,
as it is possible that instructions that were rejected earlier now do match.
For example, consider the following patterns:
.DS
.ft 5
.ft CW
cmp $0, X -> tst X ;
mov REG,X : tst X -> move REG.X ; /* redundant test */
.ft R
.DE
If the input is:
.DS
.ft 5
.ft CW
mov r0,foo : cmp $0,foo
.ft R
.DE
then the first instruction is initially rejected.
However, after the transformation
.DS
.ft 5
.ft CW
cmp $0,foo -> tst foo
.ft R
.DE
the following optimization is possible:
.DS
.ft 5
.ft CW
mov r0,foo : tst foo -> mov r0,foo
.ft R
.DE
@@ -725,7 +725,7 @@ Identifiers are sequences of letters, digits and the underscore ('_'),
beginning with a letter.
.PP
.DS
.ft 5
.ft CW
table -> {parameter_line} '%%;' {variable_declaration} '%%;'
{entry} '%%;' user_routines.
.ft R
@@ -735,7 +735,7 @@ constants, variable declarations, pattern rules and
user-supplied subroutines.
.PP
.DS
.ft 5
.ft CW
parameter_line -> identifier value ';' .
.ft R
.DE
@@ -792,7 +792,7 @@ the line is not optimized.
Optimization does, however, proceed with the rest of the input.
.PP
.DS
.ft 5
.ft CW
variable_declaration -> identifier {',' identifier} restriction ';' .
restriction -> '{' anything '}' .
@@ -816,7 +816,7 @@ Inside the expression, the name VAL stands for the part of the actual
The expression may contain calls to procedures that are defined in the
user-routines section.
.DS
.ft 5
.ft CW
entry -> pattern '->' replacement ';' .
pattern -> instruction_descr
@@ -862,7 +862,7 @@ which contains the mnemonic of the first instruction of the
rest of the input. (REST is a null-string if this mnemonic can
not be determined).
.DS
.ft 5
.ft CW
user_routines -> anything .
.ft R
.DE