How does your experience compare with this?
https://youtu.be/Cs26_b2KU_A

Haven't used SysSpeed before and it doesn't seem to run with default install, but I usually have around ~5.5MB/s fast->chip transfer. Experience doesn't seem that far from what AmiPal video is showing, but since I don't see any hard numbers (from a quick browse) it's hard to compare.

I was wondering if you have something misconfigured like not remapping the ROM or execbase etc to fast memory. Anything causing the 060 to underperform.

Amipal was getting 16fps fullscreen 1x1 in the opening location and up to 30 at 2/3 size. At least that's what the FPS counter was reporting. General gameplay at 2/3 size was around 25-30fps in his video.

Unrelated, how do we feel about having a button that disables the auto aim adjustment and maybe adds a small crosshair when it's inactive?

auto aim disable option, yes please :great

crosshair, not against it but, optional maybe? would have to try it with and without to be sure.

edit: also this option should only apply to mouse control as you would get whooped without auto aim in keyboard only and joystick control.

builds of the current main

any feed back is a bonus :)

dumb question coming,

when the sfx were moved to fast memory did this create extra lag on c2p to chip ram?

This was done ages ago, long before Amipal tested it. There's a software mixing step anyway, which is hampered by having the data in chip memory.

Does anyone actually use Joystick?

Today I am planning on doing a little refactoring. Previously I extracted a ton of zero initialised player global variables into the BSS section which now contains basically duplicate definitions for player 1 and player 2 (tiny bit of reordering needed). There's a lot of duplicated code that does the exact same thing for each player and differs only in which set of globals it refers to.

I've created a proper STRUCTURE definition to match the memory layout so that the code can be reworked to accept the appropriate player as a pointer and the same code can be used for both players. Less code to maintain and easier to add new features when you only have one copy.

It might be useful to add a versioning to the executable file/s .. in this way it would be easier to track the (beta) version used :-)

I had a quick blast on the version in post 2007 on my A1200 with ACA1230/56 and in the default smaller screen size (2/3?) it plays quite nicely. It’s quite playable with 10+ FPS most of the time.

Thanks to all involved for all your efforts with this.

How does pressing F8 affect your framerate?

We probably need to adopt semantic versioning because we each have various branches testing different stuff

I'll try and get that sorted today/tomorrow on my 1200/060. Any specific things you want me to look out for?

I remember that the original version was completely unplayable on my 1230-50 back in the day which should be something like 90% of the speed of your ACA. Overall performance seems to have improved a LOT with all the optimisations. Little drops of water, little grains of sand, make that mighty ocean and that pleasant land!

nothing in particular, just general stuff speed, any 'new' glitches, etc. but do try the 'F8' option for simple walls please :great

I have a pretty hefty PR I've tagged @pipper for review. Now that the player data has been bss'd and reordered a bit, loads of player 1/2 duplicate code has been reduced to a single path. I've playtested a few levels and have no reason at all to assume anything is broken.

What I need is a way to test 2 player mode. Can 2 instances of UAE talk to each other via some sort of virtual serial port?

Awesome stuff :great

please do check out the 'F8' option for simple walls like Karlos said, it would be nice to know how big a difference it makes :great