|
You are about to learn about endian issues. You can read all about it on the wikipedia[^]
This is an long standing issue -- and is based on the processor on your computer. For all of us Windows, we operate as "little endian" -- whereas mac and java folks operate in "big endian" format.
Endianness is just how the processor stores numbers in byte format. There's some handy functions used for socket programming (but also available generally) -- see ntohs()[^]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~<br />
Peter Weyzen<br />
Staff Engineer<br />
<A HREF="http://www.soonr.com">SoonR Inc -- PC Power delivered to your phone</A>
|
|
|
|
|
Peter Weyzen wrote: For all of us Windows, we operate as "little endian" -- whereas mac and java folks operate in "big endian" format.
Endianness is just how the processor stores numbers...
So if the processor ultimately determines how the numbers are stored, what difference does the OS or the programming language make?
"Normal is getting dressed in clothes that you buy for work and driving through traffic in a car that you are still paying for, in order to get to the job you need to pay for the clothes and the car and the house you leave vacant all day so you can afford to live in it." - Ellen Goodman
"To have a respect for ourselves guides our morals; to have deference for others governs our manners." - Laurence Sterne
|
|
|
|
|
it was just a generalization -- about how windows runs on intel, and intel is little endian
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~<br />
Peter Weyzen<br />
Staff Engineer<br />
<A HREF="http://www.soonr.com">SoonR Inc -- PC Power delivered to your phone</A>
|
|
|
|
|
IMHO the above statement is clear:
values are stored inside the file as two bytes integers, with most significative byte first, for instance 1000 (that is 03E8h in hexadecimal notation) is stored as follows:
first byte = 3 (03h); second byte = 232 (E8h)
that is:
3 * 256 + 232 = 768 + 232 = 1000</code>.
On the other hand, negative numbers are stored using Sign and magnitude representation (see [^]), hence -1000 will become 83E8h in hexadecimal notation ( 8h=1000binary, i.e. sign bit ON).
Whenever you need to retrieve values you have to always check the sign bit. I give a C code snippet (error checking is left as an exercise to the reader... ):
unsigned char a,b;
int value;
fread(&a, sizeof(a), 1, fd);
fread(&b, sizeof(b), 1, fd);
value = ((a & 0x7F) << 8 ) | b;
if ( a & 0x80)
{
value = -value;
}
Hope that helps
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
|
|
|
|
|
Thanks a Lot CPallini , by the way your mentioned code can go in error state ???
|
|
|
|
|
Well, you have always to check the return value of the fread function.
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
|
|
|
|
|
Hi All,
I am using MFC Activex Control project in VS 2005.
In my project I am doing something like hooking. I have set these hooks from application.In this hooking , I have created some windows procedure(using CALLBACK function) through which I get messages into it contineously(Messages are like when my desktop window changed,cursor position changed,anything something happens on the the screen).For this I have to check contineously for update.
And now my problem is like that when I call this hooking function than it starts executing and it does well and here I get all tese messages perfactly but during these constant message transfering my cpu usage goes upto 50%.So, that is very bad for me I want to reduse this cpu overhead near upto 10%.So, do you have any suggetion what to do for that???
I will Appreciate your Answer.
Thanks in Advance.
Ashish Bhatt
|
|
|
|
|
Here I am using below code to contineously checking for the messages.
<br />
MSG msg;<br />
<br />
while(1)<br />
{<br />
<br />
if(!PeekMessage(&msg,m_hwnd,NULL,NULL,PM_REMOVE))<br />
{<br />
CheckUpdate();<br />
}<br />
else if(msg.message == SCREEN_UPDATE)<br />
{<br />
<br />
RECT rect;<br />
rect.left = (SHORT)LOWORD(msg.wParam);<br />
rect.top = (SHORT)HIWORD(msg.wParam);<br />
rect.right = (SHORT)LOWORD(msg.lParam);<br />
rect.bottom = (SHORT)HIWORD(msg.lParam);<br />
<br />
<br />
<br />
<br />
m_region.AddRect(rect);<br />
<br />
<br />
<br />
<br />
}<br />
else if(msg.message == MOUSE_UPDATE)<br />
{<br />
MessageBox(NULL,CString("Mouse Update Message is Received"),CString("RTMPTest"),0);<br />
}<br />
else<br />
{<br />
MessageBox(NULL,CString("Peek UP Message But Not Update Messsage"),CString("RTMPTest"),0);<br />
}<br />
}<br />
Plz do helpful to me if you have any idea.
Thanks.
Ashish Bhatt
|
|
|
|
|
while (1)
{
// peek...
}
Add a Sleep(100) as a simple start...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~<br />
Peter Weyzen<br />
Staff Engineer<br />
<A HREF="http://www.soonr.com">SoonR Inc -- PC Power delivered to your phone</A>
|
|
|
|
|
Peter Weyzen wrote: Add a Sleep(100) as a simple start...
if any message come between that sleep(100) will be lost or delayed to processed by main application, as hook is working here
"Opinions are neither right nor wrong. I cannot change your opinion. I can, however, change what influences your opinion." - David Crow
Never mind - my own stupidity is the source of every "problem" - Mixture
cheers,
Alok Gupta
VC Forum Q&A :- I/ IV
Support CRY- Child Relief and You
|
|
|
|
|
So, if you have any other suggetion then you can share it.
Ashish Bhatt
|
|
|
|
|
Thanks for reply.
i used sleep() before Peekmessage() and it works.But is there no any other way?.
And also I want to ask one question that is it right way to check current process cpu usage from task manager?? because I have heard that it is not perfact way.
Thanks.
Ashish Bhatt
|
|
|
|
|
Did you see my suggestion ? What about using GetMessage instead of PeekMessage ?
|
|
|
|
|
Use GetMessage instead of PeekMessage. This way, the GetMessage will block until a message is received in the queue. And you won't need to poll for messages (which consumes CPU cycles).
Also, you should check for the WM_QUIT message (and exit your infinite loop) otherwise your application will never terminate.
|
|
|
|
|
If you're on a dual-core machine -- this 50% is likely 100% for one processor -- which is bad!
Have you confirmed this in the debugger? If one of your threads is running at this rate, it should be easy to do a "break all" in the debugger and see what the threads are up to....
If you have a loop that's just going crazy -- you could just stick a sleep in there...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~<br />
Peter Weyzen<br />
Staff Engineer<br />
<A HREF="http://www.soonr.com">SoonR Inc -- PC Power delivered to your phone</A>
|
|
|
|
|
Hi All,
Can any one please tell me this. I have a MDI window, i am opening a child window on it where i am doing some gdi rendering after reading data from a file.
Just after reading data from a file and rendering it on the child screen, memory uasage in my task manager is ahowing 37K. Now I am minimizing the application and memory usage dorps to 2K. Now if i miximize the application it is occuping only 5K.
Can any one please give me some clue what might the reason.
With Thanks in advanced
Sandip
|
|
|
|
|
|
this is any solution for this
Sandip
|
|
|
|
|
Perhaps this article[^] can give you an explanation.
"It's supposed to be hard, otherwise anybody could do it!" - selfquote
"High speed never compensates for wrong direction!" - unknown
|
|
|
|
|
I always get y_num=0 when y_num=65536
How to get rid of this problem (variable type,...)
Please help.
#include<br />
#include<br />
#include<br />
#include<br />
#include<br />
<br />
typedef unsigned int WORD;<br />
typedef WORD* WORDPTR;<br />
<br />
WORDPTR *S;<br />
WORDPTR D;<br />
<br />
WORD x_num;<br />
WORD y_num;<br />
<br />
WORD a;<br />
WORD b;<br />
WORD i, j;<br />
int main()<br />
{<br />
puts("------------------ Start --------------------");<br />
puts("");<br />
<br />
a = 0;<br />
printf("Enter a = ");<br />
scanf("%u", &a);<br />
<br />
<br />
b = 0;<br />
printf("Enter b = ");<br />
scanf("%u", &b);<br />
<br />
x_num = 1 << a;<br />
y_num = 1 << b;<br />
<br />
S = (WORDPTR *)malloc(x_num*sizeof(WORDPTR));<br />
if (S == NULL)<br />
{<br />
puts("Error 1!");<br />
return 1;<br />
}<br />
for (i = 0; i < x_num; i++)<br />
{<br />
S[i] = (WORDPTR)malloc(y_num*sizeof(WORD));<br />
if (S[i] == NULL)<br />
{<br />
puts("Error 2!");<br />
return 1;<br />
}<br />
}<br />
<br />
D = (WORDPTR)malloc(y_num*sizeof(WORD));<br />
if (D == NULL)<br />
{<br />
puts("Error 3!");<br />
return 1;<br />
}<br />
puts("OK!");<br />
<br />
return (0);<br />
}
-- modified at 4:15 Wednesday 14th November, 2007
|
|
|
|
|
ndtoan13 wrote: I know 1 << 256 == 2^256 = 65536
2^16 == 65536
1 << 16 == 65536
|
|
|
|
|
ndtoan13 wrote: I know 1 << 256 == 2^256 = 65536
The above is not true, in fact:
1 << 256 = 2^256 = <big>HUGE number ( ~ 10^77 ) </big>
even an hypothetical 256-bit wide integer cannot hold this number.
So I suggest to revise you convinctions  (BTW 65536 = 2^16 ).
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
|
|
|
|
|
Sorry I am wrong. 65536 = 2^16
I need to print an array[256x65536]
Thanks.
|
|
|
|
|
I know 1 << 256 = 2^256 = HUGE number ( ~ 10^77 )
It is my wrong type after 18h coding.
I want a code to print an array [256x65536]
--> x_num=256 and y_num=65536
Event I assign static value like y_num=65536, y_num still = 0
What type of this variable need to solve this problem?
Please help.
|
|
|
|
|
You need a 32-bit wide integer. Usually int it is enough, however it is system-dependent, you can check the wideness of your datatype using the sizeof operator, you need:
sizeof(you_integer_datatype) >= 4
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
