Compileable version of Alien Breed 3D II - The Killing Grounds

[AlphaAmiga]

Quickly ran the cgxbenchmark test again over my Mediator TX+.

Pretty much exactly same results for the Register to V Ram, Fast to V ram and the WritePixelArray(), however some slight improvements in the ScalePixelArray() with FPS / MB/s now coming out at;

Screen depth 15 = 13.4 / 3.0
Screen depth 16 = 10.5 / 1.5
Screen depth 24 = 5.5 / 1.2

I dont know how much this is effected by CPU speed but this was ran with the system using a TF1260 with 128MB Ram and a Rev 1 clocked at 50MHz.

I will try and get around to doing some more testing with the latest build when I can, especially now as I have sourced a Rev 6 060 for my towered A1200 :-) (just got to wait for it to arrive)

Cheers everyone, keep up the great work on this !

[abu_the_monkey]

@AlphaAmiga

good info, thanks

wish I could get the A4000 to work reliably to run some tests, one day I will have some time to give it a full strip and clean

[pipper]

I have been deep into hardware recently; building new daughter board(s) for my A3000, the Matze A3000 DB with integrated I2C and the Firebird DB with PCI slots as well as another BFG (with an 040 this time). I have also been busy trying to get a defective Piccolo 2.1 graphics card fixed and ended up reverse engineering most of the pcb into schematics (hey, anyone out there with the same card who likes to help out?)
I'm more busy in RL these days. Doing HW stuff has been easier for me as it required less being "in the the zone" and I can get quickly in and out of it. I have not spent time with AB3D recently unfortunately. But I'm closely following whats going on here :-) Keep on hacking!

[Karlos]

Originally Posted by pipper

I have been deep into hardware recently; building new daughter board(s) for my A3000, the Matze A3000 DB with integrated I2C and the Firebird DB with PCI slots as well as another BFG (with an 040 this time). I have also been busy trying to get a defective Piccolo 2.1 graphics card fixed and ended up reverse engineering most of the pcb into schematics (hey, anyone out there with the same card who likes to help out?)
I'm more busy in RL these days. Doing HW stuff has been easier for me as it required less being "in the the zone" and I can get quickly in and out of it. I have not spent time with AB3D recently unfortunately. But I'm closely following whats going on here :-) Keep on hacking!

It's all good. It's now possible (not merged yet) to build the "release" and "dev" versions of the engine, the latter with the various metrics and experimental stuff included.

[Karlos]

This PR is turning into a general bunch of chores, but doing them has led to the identification and removal of more dead code

Which was nice...

[abu_the_monkey]

nice

[Karlos]

Originally Posted by abu_the_monkey

nice

It's revealed quite a bit more that I'll get rid of before merging as a maintenance update.

[Karlos]

I'm going to merge my dev build and chores PR.

After this you can make, to get the release version, or make dev to get the instrumented developer version.

The chores have been tidying up the animation code a bit, lots of cruft in there.

I also think there's a potentially bad buffer overrun (but it's read only not write) in the lighting animation tables but I'll need to confirm that with a test level. It may explain why areas randomly light occasionally up for no apparent reason. The animation pointer table has entities for the 7 defined animation types only, but the editing tools allow up to 16 to be used. Those will probably run off into code memory and only stop when hitting the 999 magic value.

I propose extending the tables to make sure there are 16 valid entries, but it would be awesome if they could be loaded as mod resources. One for the Todo list.

[Karlos]

Squash merged.

[abu_the_monkey]

top stuff

anyone know of a newer version of make for amiga os than 3.76?
I don't use the cross dev stuff and make 3.76 it don't like the makefile (which ain't a big deal as I just made individual makefiles and run scripts)

it would be nice if it worked as intended without modifying but like I said not problem either way.

[Karlos]

Originally Posted by abu_the_monkey

top stuff

anyone know of a newer version of make for amiga os than 3.76?
I don't use the cross dev stuff and make 3.76 it don't like the makefile (which ain't a big deal as I just made individual makefiles and run scripts)

it would be nice if it worked as intended without modifying but like I said not problem either way.

Maybe an Amiga make compatible makefile can be added as well.

[abu_the_monkey]

I don't think it is needed.
most people ain't going to build this on a real amiga, they will use the cross dev stuff.
I just haven't set it up cos I haven't felt the need.

[Karlos]

Originally Posted by abu_the_monkey

I don't think it is needed.
most people ain't going to build this on a real amiga, they will use the cross dev stuff.
I just haven't set it up cos I haven't felt the need.

I'm running entirely "virtual" at the moment so cross building all the way, lol.

[Karlos]

I should point out that this PR isn't the one I intend initially but I had a ton of outstanding bits that just seemed to be lying around lol.

[abu_the_monkey]

I thought as much, the code is becoming much cleaner though

[Karlos]

Originally Posted by abu_the_monkey

I don't think it is needed.
most people ain't going to build this on a real amiga, they will use the cross dev stuff.

"Most of the people" building this are in this thread. Discussing this point

[abu_the_monkey]

yeah what I meant was, I am that guy that does everything the hard way when there are obvious better solutions, but, sometimes it pays off

[Karlos]

Just a sneak preview of something I'm working on. I literally just got this working 10 mins ago...
[ Show youtube player ]

So far this is just the render time (no c2p or vsync) per frame. The effect of the door overdraw is pretty obvious here...

[abu_the_monkey]

looking good

[Karlos]

Originally Posted by abu_the_monkey

looking good

I was thinking to make it a solid line but I actually think it's less distracting as a plot. It uses a round-robin approach so there's no actual shuffling of values or anything to get the scrolling