new amiga 500 doom like !

[Mathesar]

Originally Posted by malko

Even if I showed the "on work" video to my mini-Godzillas that are a bit bigger, I would not put a PEGI "6" on the box for that kind of game .

True, but I think the few red pixels on the enemies didn't disturb him I. Now, the more detailed 4K games like Doom Eternal on the other hand... I will definately not let him near those!

[d4rk3lf]

I watched Conan Barbarian when I was 7 years old, and loved it so much.
Should I be worried?

[rothers]

Originally Posted by d4rk3lf

But Dread is not a raycaster.

OK then, an engine which is not true 3D, but built from a bsp tree or whatever.

[KK/Altair]

can you write more about your C2P? e.g how textures are organized, blitter passes etc.

On Atari ST, C2P works this way:
1. Fetch a word of chunky data (2 logical pixels, 1 pixel per byte).
2. Convert it to planar with LUT1 lookup.
3. Fetch second word of chunky data (another 2 logical pixels).
4. Convert it to planar with LUT2 lookup, ORing the LUT results together.
5. MOVEP the result to screen.
Repeat for every 8 screen pixels.


One logical pixel is 2 screen pixels, so the above is converting 4 logical pixels to screen.
Everything is heavily unrolled. I discussed the above already with ST coders and found a way to save 8 cycles, which I'll implement in the future.


The every second scanline is then duplicated with CPU (on ST, using register-heavy MOVEM.L) or STe Blitter.

Though i wish sensitivity would go to 20 it is totally playable.

Your wish will be honored.
I didn't expect original mouses to be that tanky.
(I'm using self-made mouse converted from PC)

Me too. Too slow and I don't like this kind of controls. Doesn't fit the Wolf-type of game imo.

You need capability to walk in any direction freely while aim in any possible other direction to play on higher difficulties.

I think I saw KK use a non-Amiga mouse on his A500 (Maybe PS/2 via adapter?) that is probably more sensitive.

I have Amiga mouse, but didn't tried it because it looked terrible (dirty and completely yellow).
The mouse I'm using WAS a PS/2 mouse, but I unsoldered the chip inside, soldered wires directly to the ball optics and buttons and made it Amiga-compatible.

PS: on the youtube comments some people are moaning about how this is no "real" doom, that they are "not impressed" and how this should have been "done on an Archimedes" instead. If Dread manages to provoke these kind of reactions... then it actually means it must be *really* good!

Such people still amaze me - what is the point of wasting energy on something you don't enjoy?

PS3: @KK you need to clean you Amiga 500

Yeah, probably.
At least I cleaned my desk!

I tested on Amiga 600 (2M chip + 4M fast) with Commodore 1084s crt, and here the fluidity of the movements has a noticeable boost. Much more than I thought.

Dread loves Fast RAM. All the stuff is done on the CPU while Blitter does chunky 2 planar in the background (interrupts!).
Having enough Fast essentially means that you have chunky 2 planar for free, as Blitter no longer slows down the CPU.

Apart from 14Mhz, probably 1MB chip and 7 MB fast Ram also helps a lot.

Yeah, definitely that.
It's a far shot from the stock A500. Much closer to the A1200.

Wow, runs great on my 1200 and managed to get it to work on my cdtv, slight issue though as it will only run if i turn off the cd drive (frees about 80k memory) but my cdtv does have 2mb chip and 8mb fast so this is slightly odd - maybe the game is using a specific area of memory?

Nope. Right now the game is a pure EXE loaded via Startup-Sequence. The system loads it.

Ask your kickstart why has he failed to fit everything in a space so vast.

Raycasters have loads of limitations, however they are super for Doom style games and they can be really fast, as shown by Dread.

Two corrections:
1. Dread is not a raycaster. It renders polygons.
2. Raycasters are crap. Sorry.



Ad. 2. My Wildcat demo featured a Wolf3D-like sequence.
The first implementation used raycasting optimized to death in asm.
The second implementation tried using polygons and the first prototype in C got the same level of performance. Then it only got better.
Raycasting is fun, because it takes care of all the visibility problems, but casting rays for every screen column is just slow.
Polygons tell you immediately what to draw and where to draw, so they are simply faster.
The only issue is that you have to decide which polys are visible, but that can be precomputed.

I was running the A500 stock except for the CPU which is an 68010. Would that make any difference?

68010 is mostly a system architecture upgrade, allowing better system-user separation (e.g. allowing WHDLoad stay in the background without being killed by the game it ran).
The 2-instruction "cache" was in my opinion quite pointless, especially that in most cases MOVEM does much better job at filling/copying larger regions.

PS: My 6 year old son loves it and completed the level in "normal" setting

How do you both feel about the blood and death animations of the enemies?
Asking personally. I haven't yet shown these animations to my own 6 year old son.

Now, the more detailed 4K games like Doom Eternal on the other hand... I will definately not let him near those!

Yeah, IMHO they came a bit too far with that.
I love all things Doom, but didn't play the never ones just because I'm not a fan of watching all the gory death animations shown in my face again again.
Completely pointless, from the gameplay's perspective.

[grond]

Originally Posted by KK/Altair

On Atari ST, C2P works this way:
1. Fetch a word of chunky data (2 logical pixels, 1 pixel per byte).
2. Convert it to planar with LUT1 lookup.
3. Fetch second word of chunky data (another 2 logical pixels).
4. Convert it to planar with LUT2 lookup, ORing the LUT results together.
5. MOVEP the result to screen.
Repeat for every 8 screen pixels.

Oh wow, one MOVEP for every 8 pixels, I wonder how this performs on an 060...

[Cyprian]

Originally Posted by KK/Altair

On Atari ST, C2P works this way:
1. Fetch a word of chunky data (2 logical pixels, 1 pixel per byte).
2. Convert it to planar with LUT1 lookup.
3. Fetch second word of chunky data (another 2 logical pixels).
4. Convert it to planar with LUT2 lookup, ORing the LUT results together.
5. MOVEP the result to screen.
Repeat for every 8 screen pixels.


One logical pixel is 2 screen pixels, so the above is converting 4 logical pixels to screen.
Everything is heavily unrolled. I discussed the above already with ST coders and found a way to save 8 cycles, which I'll implement in the future.


The every second scanline is then duplicated with CPU (on ST, using register-heavy MOVEM.L) or STe Blitter.

thanks

actually I debugged and analyzed that in Steem Debugger.


now I'm just sitting in front of the WinUae Debugger.
But anyway what about Amiga? besides 6bit pixels on the ST vs 8bit on the Amiga's version, are there any other differences?
And how C2P looks like? Is it pure blitter or mixed with the cpu? Can you please describe C2P steps?

[KK/Altair]

But anyway what about Amiga? besides 6bit pixels on the ST vs 8bit on the Amiga's version, are there any other differences?
And how C2P looks like? Is it pure blitter or mixed with the cpu? Can you please describe C2P steps?

On Amiga, C2P is done purely on Blitter. Blitter does two full passes extracting bit groups from chunky and putting them together for planar. In each step bits from two sources are mixed and interleaved.

[S0ulA55a551n]

Originally Posted by KK/Altair

On Atari ST, C2P works this way:

Yeah, IMHO they came a bit too far with that.
I love all things Doom, but didn't play the never ones just because I'm not a fan of watching all the gory death animations shown in my face again again.
Completely pointless, from the gameplay's perspective.

You might not like the death animations, but doom eternal is a masterpiece, ands especially from a gameplay point of view

[Cyprian]

Originally Posted by KK/Altair

On Amiga, C2P is done purely on Blitter. Blitter does two full passes extracting bit groups from chunky and putting them together for planar. In each step bits from two sources are mixed and interleaved.

on the Atari you use 6bit pixels chunky (actually 8bit but two unused), and you translate them with an intermediate table into 4bitplanes. Thats clear for me.
But somewhere I read that on Amiga you use all 8bit per chunky pixel. Is it true?
Does it mean you translate that 8bit pixel with the blitter into 4 bitplanes? All this in these two passes?

[BSzili]

Originally Posted by KK/Altair

Such people still amaze me - what is the point of wasting energy on something you don't enjoy?

You're gonna get that no matter what you do, you can't please all the people all the time. There will always be someone who feels "left out", especially since you achieved something no-one thought was possible

[KK/Altair]

But somewhere I read that on Amiga you use all 8bit per chunky pixel. Is it true?
Does it mean you translate that 8bit pixel with the blitter into 4 bitplanes? All this in these two passes?

Yes, exactly that. The format of a logical pixel is 'aabbccdd'. Chunky buffer is made of interleaved columns, so next byte is already 8 screen pixels to the right. That's 4 screen pixels already in the correct position. And each pass doubles that number.

[d4rk3lf]

BTW,
KK released on Discord, Dread tools for creating levels.

Grab it here:

The Dread mapmaking SDK ver. 1907 links:
- the SDK download - http://devkk.net/files/dread/Dread-M...g-SDK-1907.zip
- documentation wiki - http://devkk.net/wiki/index.php/Dreadtool)

You're still not be able to load it in the engine, but you can preview it on PC at least, and prepare your levels, until complete moding tools are released.
There's nice discussion about Dread on Discord.

[KK/Altair]

You might not like the death animations, but doom eternal is a masterpiece, ands especially from a gameplay point of view

That much I gathered and would love to play it.
I only would love to be able to skip the watching the same ridiculously gory animations playing right in my face and just focus on the game.
If there's a mod for that, I'm in.

[Foebane]

Originally Posted by S0ulA55a551n

You might not like the death animations, but doom eternal is a masterpiece, ands especially from a gameplay point of view

Doom is meant to be run and gun, not the stupidly complicated crap they put into Doom Eternal. Must be why I liked Doom 2016 and love Doom 3 most of all, they're next-gen engines that retain the core simple gameplay.

Doom Eternal was more of an RPG, almost, with the ridiculous amount of resource management and weapon upgrades, not to mention those annoying mechanics that were absent in Doom 2016, like monkey bars, wall-climbing, those damned collapsing floating things, and the double-dash, as well as the Mario-style layouts and colourful bobbing pickups. And as for the Marauder, that bastard almost made me throw my PC out of the window. As it is, that fate was reserved for my GTX 960 (yes, I threw it out the window).

[Retro1234]

Wow a playable demo has been released for both Amiga and ST
Amazing I don't think anyone can believe it.

[Tigerskunk]

Originally Posted by BSzili

You're gonna get that no matter what you do, you can't please all the people all the time. There will always be someone who feels "left out", especially since you achieved something no-one thought was possible

That bloke on youtube is its own brand of nutjobness, though.

[malko]

^ I find this comment funny though

[gimbal]

I do believe it was John Romero who said that in an interview, but whatever...

[utri007]

Who childish adults can be, surprices me all the time.

[d4rk3lf]

Originally Posted by gimbal

I do believe it was John Romero who said that in an interview, but whatever...

It was Carmack reply on the mail, where he stated that not even 040 is able to run Doom properly without chunky mode.
https://amitopia.com/wp-content/uplo...nt-720x720.jpg

That statement is simply wrong, as Doom run really nicely on 040, and on 020/030 it runs no worse then on 386 (I think it runs even better).

When Dread introduce stairs, it will be full slap in the Carmack face