Could you test this ?

Hi !
Just an scrolling test. sorry can't post source, just executable (third party code inside)

when y=100, a lot of scrolling squares are displayed to the screen.

Could you test this program and return the displayed 'FPS', value of 'count' and your video card/config info. Many thanks !


Press escape to quit the program.

http://www.bayre.com/zigwigwis/scroll.exe

.

12 FPS at row 1601.
Radion 9250,768 MB ram, CeleronR CPU 2.93 GHz.

ok thanks for your help

unless the slow down is deliberate i'd say that routine is broken.

yes need fix some bugs...

if froze after about a 30 seconds than it crashed.

P4 Laptop 2ghz 768MB RAM Mobility redeon 7500
Windows xp sp2

i'll post a corrected version soon.

80 FPS, Count between 2 and 3
Crashes after like 15s
Will test with the new version once again.

Core Duo T2500, 7600Go 256MB, 2GB RAM

Hovered around 750 FPS, Count satyed on 1.
Spec:
AMD Athlon 64 3500+ 2.2Ghz, 1.5 Gb RAM, GeForce 6800GT 256Mb

the new version is online.

650FPS, P4 3.2GHz, 1GB, Radeon 9800XT, Win XP Home SP2.

Still say it will look nice with flip 1 unless you are just using Flip 0 to determine the max FPS of course.

You still need to fix the issue with the jiggling tiles that occurs after they wrap fromthe bottom to the top. You are setting the 6 coord to an absolute value instead of subtracting a screen's worth of tiles so that it retains any fractional values after the decimal point, this should stop the jiggling on the "recycled" tiles.

thanks Grey. i don't understand very well what's you're talking. Coz : my bad english :-< Need to find someone to translate your sentences into french language. But many thanks to post here your fix idea.

FPS Max=2480 aprox
FPS Min=1430 aprox
Pentium core duo extreme 3GHz. Nvidia 6800. XP SP2

While progress the scroll the FPS is decreased.
If the scroll is plus large the FPS are unusable.
Probabily leak in gestion of resources: Excesive paints, memory, etc.

Bye,
Paposo

hub: Sorry I only know where the problem is in your code but not how to fix it as I don't know the correct value to subtract from the cell's Y coordinate when it wraps from the bottom to the top, you should know this value though. You must subtract it from the cells Y cooord not set the Y coord to a fixed value. Anyway to be honest the way you are doing the scrolling isn't the best way. I'll make a scrolling example and include it in the framework soon.

@grey : I understand that it isn't the best solution. This structure is to keep some compatibility with my game editor and my previous game code. I don't know why but i've some difficulty to obtain an easy implementation. Perhaps tired at the end of the year !

my english is limited so it's difficult to me to explain my choice.

I need a virtual not drawing screen before and another after the main drawn screen : to progressive draw portions of large obstacles and check collision (see picture from game and editor :)



The editor use x,y to reference a cell.

I see, but this can still be achieved without changing the coordinates of each cell and having a global ScrollY coordinate instead. You can decide to draw tiles before and after the visible screen onto the collision map, and limit it so the whole level isn't added.

Into the game it's usefull to have these coords computed with the trues values. They are used for example if the cell is a cannon. To have the coord where the cannon begin to shoot. Same as for the other objects of the game : power, missiles, rabbits, bricks. Also it usefull to check collisions.



yes, it's what's my code do (draw from Lin= TopLine To BottomLine). The level is finished when the last map line is drawing at the bottom of the visible screen. into the demo the scrolling have 100 lines, you can pass it to 500 lines. Just the two virtuals screens and the viewed screen was displayed.

Thanks to discuss about this Grey. it's usefull to me to explain my choice to see where are my errors and how to improve or modified the algorithm. hope to see your scrolling example soon. When i modified my previous code, i also modified the tiles access, via array direct access. Before 'lists' was used. But my game lost time to find the good tiles into long list (for next loop) !!!
i appreciate your work with your game framework and your help.

Hub: My example is in the showcase forum now, have a look there. The code will be released in January.

Yes it sounds like an array is much better than a list for this sort of thing.