The game seems to run slowly on multiple computers.

I've noticed this before myself, usually (but not always) it has to do with whether or not I have active objects with lots of alpha channel stuff or way too many fastloops going on.

it don't semm to be slow on my computer, but it's difficult to know the difference with yours
you should add a count showing the rateframe and then we will be able to compare

Do you have the game supporting a certain version of DirectX? Try different ones, build and test. Display Options > Display Mode > Direct3d 9 might give you the widest consistent input response and framerate.

Do you have Reduce CPU Usage toggled off or on? Build and test.

Compressing the runtime?

Do you have the overall game running at a certain frame rate? Or is each level set differently? Highly suggest they all match the same frame rate.

Do you have Machine-independent speed toggled off or on?

V-sync toggled on? Try seeing if the Window options being changed help or hurt your build.

Graphic mode set to 16 Million colors? (even if doing retro looking games, Windows can misinterpret things with the runtime with lower color palette limits and switching to it and how you might have full screen and windowed options setup will just add unneeded complications) Highly suggest staying with 16MM.

Try limiting the amount of objects per Frame/level to be lower (Specific Frame > Runtime Options > Number of objects). Example, one of my current projects LOOTRONS 1982, can sometimes have a few thousands objects per frame with their own animations and events going on. I limited them per frame that has gameplay, set to 2,600 and instantly got a better framerate. Splash logo screen, tutorial, credits, and title screens 300 objects or less. Your results may vary.

Also, any objects that reach a set pixel distance from the edge of our screen/level/window you should have to be destroyed if not needed, and any objects that leave the frame area/window get destroyed too. Sometimes weird collision bugs can happen and objects can be flying off and still be active without you seeing them off screen.

Be careful with any timers with any combos mentioned above. There can be lots of trial and error, depending on your game. Sometimes it is often an easy setting we "doom clicked" or chose by accident and forgot.

Neat little project, I added it on itch.io

it's the same speed on my old computer

Usually slowness appears when you use lots of objects in a frame.

I see in your game blocks of identical objects. If it's individual backdrop objects, replace them by quick backdrop objects in "motif" mode to reduce the number of backdrop objects (if you've 2..5+ there is a command that can do it more or less automatically in the Arrange menu in the Frame editor, but I think you haven't 2.5+ yet as you were talking about it in another thread). If it's active objects instead of backdrop objects, then maybe you could replace blocks of 4, 9 or 16 objects by a single object, 2, 3 or 4 times bigger.

Otherwise 2.5+ can optimize your game:

- the Optimize Events option automatically optimizes certain events.
- using child events also may allow you to avoid repeating slow conditions (note: not obvious though, sometimes child events may be slower, especially with For Each loops).
- the profiler can also help you to determine if some events are causing the slowness.

(sorry if this sounds like a forced ad for 2.5+

Quote from Yves

If it's individual backdrop objects, replace them by quick backdrop objects in "motif" mode to reduce the number of backdrop objects

Does this actually make a substantial difference though?

I only ask because I was struggling with terrible performance just today, which I eventually traced to a Quick Backdrop using the "Motif" fill type.
It was 500x500 pixels in size, with a 5x5 pixel motif, so equivalent to 10,000 individual tiles - and unfortunately, with much the same impact on performance!

Using the software display mode, there was no problem. Using a single 500x500 image, there was no problem. It just seems there's something very inefficient about the way "Motif" quick backdrops are drawn in D3D9.

Perhaps this has already been fixed in 2.5+ though?

Here's a quick test:
Please login to see this link.
I get 35FPS with a single "motif" quick backdrop, and 33FPS with 10,000 individual backdrops. Surely it should make more of a difference than that?

MuddyMole I have a shader that replaces the motif mode and improved performance by quite a lot, also allows you to create "quick backdrops" during runtime. I'll be releasing it soon.

5400% improvement on DX9
72% improvement on DX11
(Test done in a 8K resolution app)

Quote from MuddyMole

It just seems there's something very inefficient about the way "Motif" quick backdrops are drawn in D3D9.

Yes, it's a bit better than the same number of backdrop objects, but still not good on DX9. Much better on DX11 (you example runs at 85 fps on my machine with DX9 and 1000 with DX11).

Ah, so it has been fixed in DX11 then
I appreciate it's not much of a problem anyway, since there are easy workarounds, like creating a larger motif in whatever graphics app - not the full size of the backdrop, but maybe 100x100 or whatever (there must be a sweet-spot). It's something to bear in mind though.

Really, it was just frustrating because I didn't expect it to have such an impact. I was making a pathfinding example yesterday, which is obviously quite complicated and computationally expensive anyway, so I initially put the poor performance down to that - maybe I was running a fastloop too many times or something? I tried everything I could to optimise performance, reducing the number of objects and the size of the search area, but nothing helped. It was only when I was ready to give up and start again, and I deleted every single event - and it STILL only ran at the same speed (with <10 objects on screen, and no events) - that I realized there was something else going on. Knowing the kind of complicated effects you can do with pixel shaders etc, it just never occurred to me that something so simple as a tiled motif could be any issue at all with HWA.

NaitorStudios - Thanks, that sounds interesting. As I say, it's not even a problem - it's purely that it was so unexpected, I ended up wasting a lot of time before I was able to trace the source of the issue (which then took seconds to fix!).