+,-,*,/,**(加/减/乘/除/幂运算),乘法运算的结果的位宽是乘数和被乘数位宽的和。
在进行整数的除法运算时,结果要略去小数部分,只取整数部分;而进行取模运算时(%,亦称作求余运算符)结果的符号位采用模运算符中第一个操作数的符号。
-10%3 =-1 11%-3 结果为2
注意:在进行算术运算时,如果某一个操作数有不确定的值x,则整个结果也为不确定值x。下面的代码演示了算术运算符的操作,其结果为:
如果不指定reg signed类型,负数的运算会被当作无符号数运算,从而得到错误的结果。
Compiler version M-2017.03-SP2-11; Runtime version M-2017.03-SP2-11; Dec 28 09:33 2018
32 + 15 = 47
32 - 15 = 17
15 - 32 = -17
32 * 15 = 480
32 * -32 = -1024
32 / 15 = 2
-32 mod 15 = -2
32 mod -3 = 2
3 ** 2 = 9
32 + 15 = 47
32 - 15 = 17
15 - 32 = 4294967279
32 * 15 = 480
32 * 4294967264 = 4294966272
32 / 15 = 2
4294967264 mod 15 = 14
32 mod 4294967293 = 32
3 ** 2 = 9
V C S S i m u l a t i o n R e p o r t
`timescale 1ns/1ns
module operator;
reg signed [31:0] a = 32;
reg signed [31:0] b = 15;
reg signed [31:0] c = -32;
reg signed [31:0] d = -3;
reg signed [31:0] e = 3;
reg signed [31:0] f = 2;
reg signed [31:0] r = 1;
reg [31:0] a1 = 32;
reg [31:0] b1 = 15;
reg [31:0] c1 = -32;
reg [31:0] d1 = -3;
reg [31:0] e1 = 3;
reg [31:0] f1 = 2;
reg [31:0] r1 = 1;
initial begin
r = a + b;
$display("%d + %d = %d",a, b, r);
r = a - b;
$display("%d - %d = %d",a, b, r);
r = b - a;
$display("%d - %d = %d",b, a, r);
r = a * b;
$display("%d * %d = %d",a, b, r);
r = a * c;
$display("%d * %d = %d",a, c, r);
r = a / b;
$display("%d / %d = %d",a, b, r);
r = c % b;
$display("%d mod %d = %d",c, b, r);
r = a % d;
$display("%d mod %d = %d",a, d, r);
r = e ** f;
$display("%d ** %d = %d",e, f, r);
r1 = a1 + b1;
$display("%d + %d = %d",a1, b1, r1);
r1 = a1 - b1;
$display("%d - %d = %d",a1, b1, r1);
r1 = b1 - a1;
$display("%d - %d = %d",b1, a1, r1);
r1 = a1 * b1;
$display("%d * %d = %d",a1, b1, r1);
r1 = a1 * c1;
$display("%d * %d = %d",a1, c1, r1);
r1 = a1 / b1;
$display("%d / %d = %d",a1, b1, r1);
r1 = c1 % b1;
$display("%d mod %d = %d",c1, b1, r1);
r1 = a1 % d1;
$display("%d mod %d = %d",a1, d1, r1);
r1 = e1 ** f1;
$display("%d ** %d = %d",e1, f1, r1);
end
endmodule
按位取反~ 、按位与 & 、按位或 | 、按位异或 ^ 、按位同或 ^~(~^),在不同长度的数据进行位运算时,系统会自动的将两个数右端对齐,位数少的操作数会在相应的高位补0。
下面位运算符的代码例子:
Compiler version M-2017.03-SP2-11; Runtime version M-2017.03-SP2-11; Dec 28 10:00 2018
0101010110101010 & 11111111 = 0000000010101010
0101010110101010 | 11111111 = 0101010111111111
~0101010110101010 1010101001010101
0101010110101010 ^ 11111111 = 0101010101010101
0101010110101010 ~^ 11111111 = 1010101010101010
0101010110101010 ^~ 11111111 = 1010101010101010
V C S S i m u l a t i o n R e p o r t
`timescale 1ns/1ns
module operator;
reg [15:0] a = 16'b0101010110101010;
reg [7:0] b = 8'hff;
reg [15:0] r = 1;
initial begin
r = a & b;
$display("%b & %b = %b",a, b, r);
r = a | b;
$display("%b | %b = %b",a, b, r);
r = ~a ;
$display("~%b %b", a, r);
r = a ^ b;
$display("%b ^ %b = %b",a, b, r);
r = a ~^ b;
$display("%b ~^ %b = %b",a, b, r);
r = a ^~ b;
$display("%b ^~ %b = %b",a, b, r);
end
endmodule
逻辑运算符用来连接条件表达式,它的返回值是1位的。逻辑运算符包括逻辑与 &&、逻辑或 ||、逻辑非 !,其中&&和||是双目运算符,其优先级别低于关系运算符,而 !高于算术运算符。
< 、 > 、 <= 、 >= 如果关系运算是假的,则返回值是0,如果声明的关系是真的,则返回值是1。 关系运算符的优先级别低于算数运算符。关系运算返回值也是1位的。
逻辑运算符,关系运算符,包括下面的等式运算符,通常都是用来作条件判断的,比如if语句的判断条件。
下面的代码演示了这三种运算的结果:
Compiler version M-2017.03-SP2-11; Runtime version M-2017.03-SP2-11; Dec 28 12:42 2018
(0 && 1) = 0
(0 || 1) = 1
(!0) = 1
(00010100 >= 01011010) = 0
(00010100 < 01011010) = 1
(00010100 == 01011010) = 0
(0 == x) = x
(0 === x) = 0
$finish called from file "", line 35.
`timescale 1ns/1ns
module operator;
reg a = 1'b0;
reg b = 1'b1;
reg r = 1'b1;
reg [7:0] c = 8'd20;
reg [7:0] d = 8'd90;
initial begin
r = (a && b);
$display("(%b && %b) = %b",a, b, r);
#10
r = (a || b);
$display("(%b || %b) = %b",a, b, r);
#10
r = !a;
$display("(!%b) = %b",a, r);
#10
r = (c>=d);
$display("(%b >= %b) = %b",c, d, r);
#10
r = (c
$display("(%b < %b) = %b",c, d, r);
#10
r = (c==d);
$display("(%b == %b) = %b",c, d, r);
#10
r = (a==1'bx);
$display("(%b == x) = %b",a, r);
#10
r = (a===1'bx);
$display("(%b === x) = %b",a, r);
#10
$finish;
end
endmodule
== 、!= 、===、!== 符号之间不能有空格。
“==”和“!=”称作逻辑等式运算符,其结果由两个操作数的值决定。由于操作数可能是x或z,其结果可能为x;
“===”和“!==”常用于case表达式的判别,又称作case等式运算符。其结果只为0和1,如果操作数中存在x和z,那么操作数必须完全相同结果才为1,否则为0。
逻辑等式运算符和case等式运算符的区别===01xz==01xz
01000010xx
10100101xx
x0010xxxxx
z0001zxxxx
逻辑移位运算符:<< 、 >> , a>>n其中a代表要进行移位的操作数,n代表要移几位。这两种移位运算都用0来填补移出的空位。算术移位运算符:<<>>, <<>>左边补MSB(最高有效位,仅对有符号数),若MSB是x或z,则填x。
Compiler version M-2017.03-SP2-11; Runtime version M-2017.03-SP2-11; Dec 28 13:17 2018
(10101001011100000001001000110100>>1) = 01010100101110000000100100011010
(10101001011100000001001000110100>>1) = 01010010111000000010010001101000
(10101001011100000001001000110100>>>1) = 01010100101110000000100100011010
(10101001011100000001001000110100<<<1) = 01010010111000000010010001101000
(10101001011100000001001000110100>>>5) = 11111101010010111000000010010001
(10101001011100000001001000110100>>5) = 00000101010010111000000010010001
$finish called from file "", line 26.
`timescale 1ns/1ns
module operator;
reg [31:0] a = 32'ha9701234;
reg signed [31:0] b = 32'ha9701234;
reg [31:0] r = 32'b1;
initial begin
r = (a >> 1);
$display("(%b>>1) = %b",a, r);
#10
r = (a << 1);
$display("(%b>>1) = %b",a, r);
#10
r = (a >>> 1);
$display("(%b>>>1) = %b",a, r);
#10
r = (a <<< 1);
$display("(%b<<<1) = %b",a, r);
#10
r = (b >>> 5);
$display("(%b>>>5) = %b",b, r);
#10
r = (b >> 5);
$display("(%b>>5) = %b",b, r);
$finish;
end
endmodule
{信号1的某几位,信号2的某几位,......信号n的某几位} 将某些信号的某些位列出来,中间用逗号分开,最后用大括号括起来表示一个整体的信号。
在位拼接的表达式中不允许存在没有指明位数的信号。
{a,b[3:0],w,3'b101} //等同于 {a,b[3],b[2],b[1],b[0],w,1b'1,1'b0,1'b1}
{4{w}} //等同于{w,w,w,w}
{b,{3{a,b}}} //等同于{b,a,b,a,b,a,b} 这里面的3、4必须是常量表达式。
这是单目运算符,也包括与、或、非运算。运算规则与位运算相似,不过是对单个运算符的每一位逐步运算,最后的运算结果是一位的二进制数。
c=&B; //意思同c=((B[0]&B[1]) &B[2] ) & B[3];
A?B:C, 如果表达式A为true(1),则结果为B,如果为false(0)结果为C。若A结果为模糊值(X或Z),则B,C都被执行,则是2驱动源,按下表输出结果,如果B,C不同,或者B,C中有x,z,则输出x。?:01xz
00xxx
1x1xx
xxxxx
zxxxx
小括号优先级最高,在不清楚其它运算符优先级的情况下,可以加小括号来保证想要的表达式执行顺序。条件运算符是从右向左关联的,其它运算符都是从左向右关联的。
Verilog中运算符优先级如下表优先级运算符操作数备注
高+, -, !, ~,&,~&,|,~|,^,~^,^~一元b=-a,b=+a,b=!a,b=~a; 红色的都是缩减运算符
**幂运算符
*, /, %
+ -二元加减运算
<> ,<<>>
< , <=, > , >=
==, !=,===,!==
&二元
^, ^~, ~^二元异或和同或
|二元
&&
||
低?:条件运算