C / C++ / MFC

Richard Andrew x646-Dec-06 13:14

6-Dec-06 13:14

I think it is better to use the Win32 function GetTickCount() .

--------------------------------

"All that is necessary for the forces of evil to win in the world is for enough good men to do nothing" -- Edmund Burke

Mark Salsbery6-Dec-06 13:24

6-Dec-06 13:24

I think GetTickCount() has the same resolution as clock(). Alot of sorting can get done in
1 millisecond.

Richard Andrew x646-Dec-06 13:33

6-Dec-06 13:33

My apologies. I'll go slink back into my hole. Smile | :)

--------------------------------

"All that is necessary for the forces of evil to win in the world is for enough good men to do nothing" -- Edmund Burke

Mark Salsbery6-Dec-06 13:45

6-Dec-06 13:45

Richie308 wrote:
I'll go slink back into my hole.

Nah....just pokin fun Poke tongue | ;-P

I wish I knew how to do this actually. I suppose you'd need to be able to get the CPU clock rate
for the machine, which processor type, how many cycles each machine code instruction used takes,
and how many iterations of those codes for the given sort. My head hurts thinking about it!

Cheers,
Mark

Mark Salsbery6-Dec-06 13:42

Re: Find Merge Sort time [modified]

6-Dec-06 13:42

It's tough to measure that small of a time span on a PC.

If your platform is Windows, you could check out GetThreadTimes() API. This will be more accurate
because time spent on other threads isn't included. I'm not sure what the true accuracy is though.
I'm pretty sure there are no clocks on a PC with enough accuracy to measure your sort time.
10ms to 55ms is about the best precision, and that's a mighty long time in CPU cycles these days.

Mark

Mark Salsbery6-Dec-06 14:01

6-Dec-06 14:01

Also

QueryPerformanceFrequency() and QueryPerformanceCounter() if your machine supports it.

-- modified at 20:07 Wednesday 6th December, 2006

*Edit* QueryPerformanceFrequency() returns 3192270000 counts/sec on my machine. That sounds
pretty accurate Smile | :)

lzzluca6-Dec-06 15:01

lzzluca

6-Dec-06 15:01

thanks to all, but i don't know where to find functions in yours answers ( as GetThreadTimes() API )! I'm using C++ but i know only C... How i have to set VC++ to use them??
p.s.: my OS is win XP

Mark Salsbery6-Dec-06 15:10

6-Dec-06 15:10

I guess you never said you were using Windows.

clock() is in the C runtime (time.h) so doesn't need Windows.

The others are part of Windows:
QueryPerformance functions and GetTickCount() are in windows.h/Kernel32.lib.

Mark

p.s. Thanks to whoever posted the QueryPerformance function link here a while back.
After rambling here about PCs not having the resolution I remembered the post and had to dig
through my favorites/bookmarks (which I now remember need a serious organizing).

David Crow6-Dec-06 17:13

David Crow

6-Dec-06 17:13

Mark Salsbery wrote:
I guess you never said you were using Windows.

Does VS2005 exist on any other platform?

"Approved Workmen Are Not Ashamed" - 2 Timothy 2:15

"Judge not by the eye but by the heart." - Native American Proverb

Mark Salsbery6-Dec-06 17:55

6-Dec-06 17:55

DavidCrow wrote:
Does VS2005 exist on any other platform?

Beats me, but I try not to assume, even though I do all the time.

David Crow6-Dec-06 17:17

David Crow

6-Dec-06 17:17

lzzluca wrote:
can merge sort be too fast to find order time by this code??).

Any sort is fast when only eight items are involved.

lzzluca wrote:
I can't enlarge array (to have a bigger order time), so how i can do to find order time?

Why exactly do you need it? A sorting algorithm's time is only relevant on a particular machine. Move to a faster machine and the performance improves, but the number of comparisons/swaps does not change.

"Approved Workmen Are Not Ashamed" - 2 Timothy 2:15

"Judge not by the eye but by the heart." - Native American Proverb

lzzluca7-Dec-06 1:51

lzzluca

7-Dec-06 1:51

I'm a student! My teacher gave me a project to do for an examination: i have to compare merge sort and quicksort, using an 8 elements array... So, seems good to have order time of merge and quick sort on the same array! Or not?

David Crow7-Dec-06 2:36

David Crow

7-Dec-06 2:36

lzzluca wrote:
So, seems good to have order time of merge and quick sort on the same array! Or not?

No, time is not a very useful number, especially when comparing sorting algorithms. Consider the result if you ran merge sort on computer A, and quicksort on computer B. Unless the two computer were identical, your results would obviously be skewed.

Surely your teacher has gone over big-O notation in the context of "benchmarking" algorithms. Those numbers do not change no matter what computer platform you use.

In any case, if you still need to time them, you'll need to explain to your teacher that the smallest resolution Windows will give you is 10ms, so both algorithms, sorting 8 items, will appear to complete in 0 seconds.

"Approved Workmen Are Not Ashamed" - 2 Timothy 2:15

"Judge not by the eye but by the heart." - Native American Proverb

S Douglas9-Dec-06 1:26

S Douglas

9-Dec-06 1:26

This isn't of a whole lot of help but more food for thought.

Time is the Simplest Thing...[^]

I'd love to help, but unfortunatley I have prior commitments monitoring the length of my grass. :Andrew Bleakley:

Naming a Proxy

Richard Andrew x646-Dec-06 11:59

6-Dec-06 11:59

What determines the Name property of a proxy server that I am developing in Win32?

For instance, the proxy settings in Internet Explorer ask for a Name and port number. Well, the port number depends upon which port I bind my listening socket to, but what determines the "Name?"

--------------------------------

"All that is necessary for the forces of evil to win in the world is for enough good men to do nothing" -- Edmund Burke

User 5838526-Dec-06 12:08

User 583852

6-Dec-06 12:08

Richie308 wrote:
but what determines the "Name?

I guess its the IP address or DNS name of the machine running the proxy

System.IO.Path.IsPathRooted() does not behave as I would expect

Richard Andrew x646-Dec-06 12:11

6-Dec-06 12:11

But what if you have more than one proxy running on a single machine?

--------------------------------

"All that is necessary for the forces of evil to win in the world is for enough good men to do nothing" -- Edmund Burke

led mike6-Dec-06 12:17

led mike

6-Dec-06 12:17

Then it must be listening on a different port not a different name.

led mike

Richard Andrew x646-Dec-06 12:18

6-Dec-06 12:18

Ahh yes! That must be it. Thanks! Big Grin | :-D

--------------------------------

"All that is necessary for the forces of evil to win in the world is for enough good men to do nothing" -- Edmund Burke