Operations on different variable types

ljj3166

Operations on different variable types [Copy link]

I was DIYing a DDS here a few days ago and encountered a very interesting problem. I wrote it here for everyone to see

1. unsigned int FREQ_MHz=0,FREQ_kHz=0,FREQ_Hz=0; long unsigned int FREQ=1000;

复制代码

Three unsigned integer variables FREQ_MHz, FREQ_kHz and FREQ_Hz and an unsigned long variable FREQ are defined. Four variables are used for calculation

1. FREQ = FREQ_MHz*1000000+FREQ_kHz*1000+FREQ_Hz;

复制代码

In the above calculation, if the value of FREQ_kHz variable exceeds 50, the calculation of FREQ_kHz*1000 will fail. For example, FREQ_MHz=3, FREQ_kHz=66, FREQ_Hz=10 The result of the operation is FREQ = 3042610. I guess the problem is caused by the wrong variable type, so I changed it to

1. FREQ = FREQ_MHz*1000000+(long)FREQ_kHz*1000+FREQ_Hz;

复制代码

Everything works fine, so the question is why the calculation result is correct without forced conversion of FREQ_MHz and FREQ_Hz? Experts, please come out and add some information. The compilation environment is CCS6

freebsder

Yes, there is one that improves the entire expression to be expressed in the longest type. It should be OK. = M*100000UL + K*1000UL + H*1UL should be written more beautifully, more uniformly, and less overbearingly and forcefully. . .

ljj3166

freebsder posted on 2015-11-17 14:39 Yes, there is a way to promote the whole expression to the longest type. It should be OK. = M*100000UL + K*1000UL + H*1UL ...

M*100000 is promoted to long, but k*1000 is still hovering in Uint16_t...

freebsder

The entire expression is evaluated as the longest type of the entire statement.

白丁

Why is the result correct if FREQ_MHz and FREQ_Hz are not converted? One of the variables in the addition is a long unsigned int, and the others will be converted accordingly. Try converting the other two variables, and see if the result is the same.

白丁

long unsigned int + unsigned int, first convert the unsigned int to long unsigned int, then perform the calculation

白丁

It's strange why the result is 3042610

huaiqiao

Your result is very strange. I ran it with CB. Here is my result:

ljj3166

The calculation result of my code should be the result of unsigned integer overflow. I will post the simulation later.

ljj3166

The calculation result of my code should be the result of unsigned integer overflow. I will post the simulation later.

huaiqiao

huaiqiao posted on 2015-11-17 17:11 Your result is very strange. I ran it with CB. Please take a look at the result:

OK, please post the picture and we can take a look. O(∩_∩)O Haha~

白丁

ljj3166 posted on 2015-11-17 17:15 The calculation result of my code should be the result of unsigned integer overflow. I will post the simulation later.

Overflow? How can such a small number overflow?

辛昕

Just tell me whether your MCU is msp430

lcofjp

16-bit compiler