Please help by testing this app!

Post your Construct-made creations!

Post » Sat Oct 27, 2012 4:07 pm

I discovered a bug in the physics object that prevents progress in one of my projects and now need your help to find a limit that works.

What is needed to test?
- You need to be able to run a window with a size of 1600x900.
- You shouldn't be scared of an application crashing to desktop.

The test application won't show much. It is a test involving the physics object, stressing the cpu a lot. Three texts show (from left to right) a number of sprite particles, a percentage value and the framerate (updated once per second).

Just run the application and wait until the percentage value raised to somewhere above 95% and then doesn't change anymore. Have a look at the framerate. I'm not interested in the highest but the lowest value there. You can exit the application with the red close button in the caption bar, or by pressing the escape key.

Please redo the test at least 4 times, if it doesn't crash all the time.

Then answer in this thread with the following informations:
1) Did it crash (and how often) or didn't it?
2) If it crashed, please provide informations about your cpu (type and speed, e.g. AMD Athlon II X2 250 @ 3 GHz) and about your gpu (type, e.g. NVidia 460 GTX)
3) If it did run at least once, please provide the lowest fps.

I'm especially interested in tests from the low end side of computers, because I'm aiming for my project to be executable on those, too. But everyone who wants to test is very welcomed of course.

One test run lasts approx. 20 seconds, so the whole test won't steal you more than 2 minutes of your time. Please help me by sacrificing those 2 minutes!

Many thanks in advance!

Download: test_cpu_stress_phsx.rar
Image
B
23
S
8
G
10
Posts: 1,820
Reputation: 8,242

Post » Sat Oct 27, 2012 4:43 pm

Ran test 4 times.

Test 1. min 251 FPS, froze at 95.3% on 556 FPS
Test 2. min 250 FPS, froze at 93% on 555 FPS
Test 3. min 250 FPS, froze at 93.3% on 557 FPS
Test 4. min 249 FPS, froze at 95.3% on 555 FPS

Pressing <ESC> exited correctly every time.

My system spec is in my sig.

zenox982012-10-27 16:43:47
If your vision so exceeds your ability, then look to something closer.
Moderator
B
120
S
28
G
68
Posts: 4,844
Reputation: 48,289

Post » Fri Nov 02, 2012 2:45 am

Thank you very much @zenox98 !

For all others: I still hope I get a dozen or more test results. It doesn't steal more than 2 minutes of your time. So please help me! (Don't let me beg on my knees )
Image
B
23
S
8
G
10
Posts: 1,820
Reputation: 8,242

Post » Fri Nov 02, 2012 7:02 am

Can you make it so the window size is smaller (my max is 1280x900), or move the text objects to the top/left of the layout so they can be read.

Anyhoo, I changed the window size to 800x600 and all I can read is the "Max Sprite Particles" text. It seems to run fine until it reaches about 1600 and after that the window is not very responsive until it reads about 1700 at which point it crashes to the desktop.

2nd run crashes at about 1200.
3rd at about 1266.
4th at about 1300.

1. It crashes every time, but you can rule out window size as a cause since it crashes with a smaller window too.
2. My specs:
cpu:1.5 ghz intel celron m
gpu:Mobile Intel(R) 915GM
3. I couldn't read to fps text (I'll try to move it into view if it's needed).

Can you give me more of an idea of what the test does? One bug that causes a crash with the physics behavior that has to be worked around is you can only create a certain number (255) of physics objects per tick. It's an arbitrary limit built into the box2d library. If you create any more than 255 physics objects per tick your game will crash to the desktop, so you have to come up with a way to spread the creation over multiple ticks.

-Happy bug hunting.
B
79
S
24
G
54
Posts: 4,746
Reputation: 40,755

Post » Fri Nov 02, 2012 9:30 pm

[QUOTE=R0J0hound] Can you make it so the window size is smaller (my max is 1280x900), or move the text objects to the top/left of the layout so they can be read.[/QUOTE] I'll do so. The reason for that specific size was that the issue is related to some kind of ratio between window size, parameter settings of the physics object, the number of physics objects and another aspect. I'll try to find a ratio for a smaller size (meaning, find the point at which it starts to fail)

[QUOTE=R0J0hound] Anyhoo, I changed the window size to 800x600 and all I can read is the "Max Sprite Particles" text. It seems to run fine until it reaches about 1600 and after that the window is not very responsive until it reads about 1700 at which point it crashes to the desktop.[/QUOTE] That's exactly the bug I'm after. I would have guessed that my cpu is a bit more powerful, cause it happens to me starting at ~2300 in 1600x900 resolution. And indeed it is. (It would be a relief, if the window size could be eliminated from the possible issues affecting this behavior)

[QUOTE=R0J0hound] Can you give me more of an idea of what the test does? One bug that causes a crash with the physics behavior that has to be worked around is you can only create a certain number (255) of physics objects per tick. It's an arbitrary limit built into the box2d library. If you create any more than 255 physics objects per tick your game will crash to the desktop, so you have to come up with a way to spread the creation over multiple ticks. [/QUOTE] I knew about this issue, but the problem here is another one. The objects are created over time, not more than 3 per tick. I'll send you a pm with details

[QUOTE=R0J0hound]-Happy bug hunting.[/QUOTE] Thank you
Image
B
23
S
8
G
10
Posts: 1,820
Reputation: 8,242

Post » Sat Nov 10, 2012 3:00 pm

Thanks to ROJOhound and his tips I now will work on a cap that may work around the bug, while still keeping the comfort of the physics behavior. Hopefully the next test will show that it works well
Image
B
23
S
8
G
10
Posts: 1,820
Reputation: 8,242

Post » Wed Nov 14, 2012 10:28 pm

test1
2584 particles
95.4%
534 fps
--------
test2
2584 particles
95.6%
534 fps
---------
cpu i5 2500k/windows7x64(4gbram)/radeon6950 cu2 (2048ram)
cpu core1 =2%
cpu core2 =2%
cpu core3 =64%
cpu core4 =41%
temp ok normal 53celcius
temp desktop no work 45celcius
B
2
Posts: 8
Reputation: 266

Post » Sat Dec 22, 2012 4:16 am

No crashes.
CPU: Intel Core i7 cpu Q 740 @ 1.73 ghz
GPU: NVIDIA Geoforce GT 435M
Lowest FPS: 125 fps

Test 1:
2584 particles
98.8%
125 fps

Test 2:
2584 particles
98.5%
125 fps

Test 3:
2584 particles
98.2%
138 fps

Test 4:
2584 particles
98.4%
140fps
B
10
S
3
G
1
Posts: 103
Reputation: 1,390


Return to Your Creations

Who is online

Users browsing this forum: No registered users and 1 guest