CPU Benchmark

[KJohn]

as for 60%, it's the same answer I gave to UEZ, and I stated this in the 1st post, but your results sounds little bigger then they should be, maybe some other high resource taker application was working on the background?

My fault! The AutoIt installation I had when I first ran this was 3.2.12.1 - I updated it to 3.3.0.0 and tried again. My results are as follows:

Test: Dual/Quad Core optimization

Your time: 23.3925 sec

π (pi) result: 3.14159265358979

Your CPU: Intel® Core2 CPU T7200 @ 2.00GHz

and

Test: Dual/Quad Core optimization

Your time: 19.4273 sec

π (pi) result: 3.14159265358979

Your CPU: Intel® Core2 CPU P8400 @ 2.26GHz

Excellent program by the way! The GUI is really impressive!

That's a massive jump in performance from 3.2.12.1. My congrats to Jon and the AutoIt team!

It is a Dell Latitude D830 with 2GB memory and I tested it on Vista x32 without SP1.

The perfomance index - base score is 3,4:

Processor Intel® Core2 Duo CPU T7300 @ 2.00GHz 4,9

Memory 2,00 GB 4,8

Graphics NVIDIA Quadro NVS 140M 3,4

Gaming graphics 1022 MB Total available graphics memory 4,4

Primary hard disk 13GB Free (110GB Total) 5,2

UEZ

Sorry for wasting your time by asking you to type it in... The problem lay in the AutoIt installation version.. Edited January 7, 2009 by Koshy John

[KJohn]

@sandin

The position of the copy results window varies with the end user's screen resolution. I've seen it on 1440x900 and 1280x800. Looking at the varying positions of that little window on the two screens, I'm assuming you are working on a 1680x1080 or higher resolution (based on the screenshots that you've posted).

[HAL9000]

Test: Single Core optimization

Your time: 68.2314 sec

p (pi) result: 3.14159265358979

Your CPU: AMD Athlon XP 2000+

Test: Dual/Quad Core optimization

Your time: 43.9196 sec

p (pi) result: 3.14159265358979

Your CPU: AMD Athlon XP 2000+

[youknowwho4eva]

$pip = 0
$pip2 = 0
For $p =  0 To 100000000
    $pi = (((1/16)^$p)*((4/((8*$p)+1))-(2/((8*$p)+4))-(1/((8*$p)+5))-(1/((8*$p)+6))))
;~ MsgBox(0,"",$pi)
$pip += $pi
If $pip = $pip2 Then ExitLoop
$pip2 = $pip
;~ MsgBox(0,"",((4*atan(1/5)) - (atan(1/239)))*4)
Next
MsgBox(0,"",$pip)

gets the same answer, I'm going to try and modify to see if i can get it to calculate more decimals. It's a lot quicker then 30 seconds.

Edit: This might be a dead end for calculating more decimals because it cuts off the calculations of $pi, there for it would only be accurate to the current decimal.

Another edit, if you unquote out the last msgbox that is quoted out, it also returns pi.

Edited January 7, 2009 by youknowwho4eva

[sandin]

@sandin

The position of the copy results window varies with the end user's screen resolution. I've seen it on 1440x900 and 1280x800. Looking at the varying positions of that little window on the two screens, I'm assuming you are working on a 1680x1080 or higher resolution (based on the screenshots that you've posted).

just tested those resolutions (1440x900 and 1280x800), and it looks the same as on 1680x1050 (you are correct, this is my screen res.), the $form3 gui should place itself on the position related to main gui ($form1), regardless of screen resolution

$width2 = 200
$height2 = 92
$form3 = GUICreate("result", $width2, $height2, $size[2]-$width2-3, $size[3]-2, $WS_POPUP, $WS_EX_MDICHILD, $Form1)

where $size = wingetpos($form1)

As for results, I'm glad it turns out to be OK, your results are very good now! Btw, I also wasn't aware of the new autoit performance jump, gz to autoit team

Edited January 7, 2009 by sandin

[sandin]

$pip = 0
$pip2 = 0
For $p =  0 To 100000000
    $pi = (((1/16)^$p)*((4/((8*$p)+1))-(2/((8*$p)+4))-(1/((8*$p)+5))-(1/((8*$p)+6))))
;~ MsgBox(0,"",$pi)
$pip += $pi
If $pip = $pip2 Then ExitLoop
$pip2 = $pip
;~ MsgBox(0,"",((4*atan(1/5)) - (atan(1/239)))*4)
Next
MsgBox(0,"",$pip)

gets the same answer, I'm going to try and modify to see if i can get it to calculate more decimals. It's a lot quicker then 30 seconds.

Edit: This might be a dead end for calculating more decimals because it cuts off the calculations of $pi, there for it would only be accurate to the current decimal.

Another edit, if you unquote out the last msgbox that is quoted out, it also returns pi.

I tried several formulas (wikipedia source), and tested each one with monitoring Task Manger performances, to see CPU usage (of both cores), and eventually I decided to use these ones in my script, though there are many other ways of calculating pi number

[KJohn]

just tested those resolutions (1440x900 and 1280x800), and it looks the same as on 1680x1050 (you are correct, this is my screen res.), the $form3 gui should place itself on the position related to main gui ($form1), regardless of screen resolution

$width2 = 200
$height2 = 92
$form3 = GUICreate("result", $width2, $height2, $size[2]-$width2-3, $size[3]-2, $WS_POPUP, $WS_EX_MDICHILD, $Form1)

where $size = wingetpos($form1)

As for results, I'm glad it turns out to be OK, your results are very good now! Btw, I also wasn't aware of the new autoit performance jump, gz to autoit team

This is how it looks on 1440x900. When I use it on the 1280x800 screen, the box moves further left and a little higher. I extrapolated on that to assume that you were running at a higher resolution and display standards have a notch at 1680 so guessed that... I'll troubleshoot your code if you are unable to replicate the problem but see if you can find a reason for this happening. (Remember to change the resolution before you run the program - just in case you left the program running and then tried changing res)

[gseller]

Test: Dual/Quad Core optimization

Your time: 19.0553 sec

π (pi) result: 3.14159265358979

Your CPU: Intel® Core2 Quad CPU Q6600 @ 2.40GHz

I guess I am not doing too bad...

[sandin]

This is how it looks on 1440x900. When I use it on the 1280x800 screen, the box moves further left and a little higher. I extrapolated on that to assume that you were running at a higher resolution and display standards have a notch at 1680 so guessed that... I'll troubleshoot your code if you are unable to replicate the problem but see if you can find a reason for this happening. (Remember to change the resolution before you run the program - just in case you left the program running and then tried changing res)

here are my screenshots:

1280x800

and 1440x900

I also tested on my old P4 PC, monitor 17" (res 1024x768), and it looks the same

I made some modifications after your first post on this thread, try downloading again, and running script again, I hope it'll turn out to be OK

[KJohn]

here are my screenshots:

1280x800

and 1440x900

I also tested on my old P4 PC, monitor 17" (res 1024x768), and it looks the same

I made some modifications after your first post on this thread, try downloading again, and running script again, I hope it'll turn out to be OK

Works now! Now if you could figure out a way to max out the CPUs... I just went through your code - the part where different code is executed depending on the number of cores... Maybe I didn't go through it thoroughly enough but I don't understand how different cores will be engaged in parallel by what seems to be a single threaded application. I watched task manager as the program was run for both single and dual core opt; and for both the CPU use was marginally higher for one core and marginally lower for the other - no noticeable difference between the different optimizations except that one takes longer than the other but I feel inclined to attribute this to the fact that different code is being executed in each case.

Maybe I haven't understood the logic but this is my preliminary impression. I'll go through it later when I have more time and get back to you...

[sandin]

Works now! Now if you could figure out a way to max out the CPUs... I just went through your code - the part where different code is executed depending on the number of cores... Maybe I didn't go through it thoroughly enough but I don't understand how different cores will be engaged in parallel by what seems to be a single threaded application. I watched task manager as the program was run for both single and dual core opt; and for both the CPU use was marginally higher for one core and marginally lower for the other - no noticeable difference between the different optimizations except that one takes longer than the other but I feel inclined to attribute this to the fact that different code is being executed in each case.

Maybe I haven't understood the logic but this is my preliminary impression. I'll go through it later when I have more time and get back to you...

great, I made mistake 1st time posting script (used $size[0] instead of $size[2] and $size[1] instead of $size[3], but it's ok now ). And ya, there is a catch, which is somehow strange, I assume you are using Firefox, so am I, and when firefox is ON while I'm running test for single core, both of the cores interact, but if only running application is Cpu benchmark, then there is difference between dual and single core test. The part in the code that switches tests with $cores = 1 or 2 or 3 or 4 is not the number of cores, and thread optimizer, it's just a switch between different PI formulas. Dunno if it is possible to make this in AutoIt (since AutoIt is single threaded tool), but if you run only CPU benchmark and monitor task manager (all other applications should be off), you should notice a difference like the 3rd screenshot in my 1st post

Anyway, tnx for the feedback, I appreciate it

[gseller]

Upgrading to 3.3.0.0 boosted my times too..

Test: Dual/Quad Core optimization

Your time: 18.3713 sec

π (pi) result: 3.14159265358979

Your CPU: Intel® Core2 Quad CPU Q6600 @ 2.40GHz

Test: Dual/Quad Core optimization

Your time: 18.4052 sec

π (pi) result: 3.14159265358979

Your CPU: Intel® Core2 Quad CPU Q6600 @ 2.40GHz

Test: Dual/Quad Core optimization

Your time: 18.4519 sec

π (pi) result: 3.14159265358979

Your CPU: Intel® Core2 Quad CPU Q6600 @ 2.40GHz

Awesome script.. Thanks!

[sandin]

I'm glad you like it

[FireFox]

@sandin

My results :

Test:   Dual/Quad Core optimization
Your time:   10.7416 sec
π (pi) result:   3.14159265358979
Your CPU:   Intel(R) Core(TM)2 Extreme Quad CPU QX9770 @ 3.2 GHz

Woosh !

Cheers, FireFox.

Edited January 8, 2009 by FireFox

[youknowwho4eva]

I have a suggestion for the modification of how you calculate pi and usage of the gdi+. If I played with gdi+ I'd do it but I haven't tried learning it yet. I'd make a square with a circle in it, I'm sure you've see the circle in a square method. Then randomly place points and calculate pi from those points. I haven't thought of a way to do the license plate method yet .

[sandin]

I'm a GDI+ beginner as well , this was my 1st time using GDI+

you mean something like this:

@sandin

My results :

Test:   Dual/Quad Core optimization
Your time:   10.7416 sec
π (pi) result:   3.14159265358979
Your CPU:   Intel(R) Core(TM)2 Extreme CPU QX9770 @ 3.2 GHz

Cheers, FireFox.

nice CPU! Edited January 7, 2009 by sandin

[youknowwho4eva]

http://www.dreamincode.net/forums/showtopic30265.htm

something like the code that guy wrote. If you have a square with a circle in it that has a diameter of the height of the square and pick random points in the square. The number of points that land in the circle divided by the total number of points all divided by 4 is pi.

[sandin]

interesting method, tnx for the link, I'm sure that way would take more CPU resources then only calculating formula (like it's doing now), but I lack GDI+ skills (as mentioned in my previous post)

Edited January 7, 2009 by sandin

[FireFox]

@sandin

Thanks ! I think I have to have good processor to script in autoit xD

Cheers, FireFox.

[youknowwho4eva]

Well you could always have it doing the math in the background. Would still eat up more processor. Hopefully someone will give us an example of it