The Atari 2600+ is live for preorders!

[tradyblix]

51 minutes ago, Brad_from_the_80s said:

Yeah, they probably wouldn't even really have to be worth a million dollars or whatever crazy amount it was ($50k? $100k?), something more modest but just unique and one-of-a-kind, or extremely limited quantities or something.  And/or also release SQ Air World on a limited cart and a new comic and make a big deal about it as the final phase of the quest.

yeah you could even tweak it and make it slightly different for the contest than the atari 50 version. 

[protomank]

14 hours ago, Ben from Plaion said:

Yeah, I think they were signed Jack Tramiel prototypes!

 

Jokes aside I got together a VIP box, Prototype PCB, poster collection and a bunch of accesories and games. 

You can call me crazy, but I got more excited for the PCB than the VIP Box for a moment. The nrd engineer inside me speaks louds sometimes

[Crimson Cross]

I got my shipping email from Atari, it'll be fun playing both old, new and the AtariAge homebrews once I get the unit in...

[Thomas Jentzsch]

5 hours ago, Stephen said:

The hell you talking about?  2600 emulators were silky smooth from the 486 days and up.  It doesn't take an i7 to emulate one - trust me.

They were cheating. Also z26 (written is Assembler!) barely run at full speed on my Pentium 100. Stella is not cheating anymore, therefore it requires significantly more CPU power. Mainly from one single core. 

[+Intellivision Master]

1 hour ago, Crimson Cross said:

I got my shipping email from Atari, it'll be fun playing both old, new and the AtariAge homebrews once I get the unit in...

I also got the shipping email from Atari.  But I didn't have it preordered.  As for the homebrews, some don't work.  Hopefully that firmware update corrects it.

[protomank]

4 hours ago, Thomas Jentzsch said:

They were cheating. Also z26 (written is Assembler!) barely run at full speed on my Pentium 100. Stella is not cheating anymore, therefore it requires significantly more CPU power. Mainly from one single core. 

Hum, I have no idea on how the Stella code is structured, but it seems this could be improved, as CPU, sound and TIA can be on their own thread, even the translation from 6507 to ARM/x86 in one, to allow multiprocessing to take advantage of multiple cores.

 

Anyway, I did not notice anything like that on my PSX Classic, that is where I do run emulators, but I will take a closer look. It is a old SoC, but the best from those mini-consoles, I do not see CPU/GPU age being, per se, the issue. (picking an old and weak, of course is).

[eCockpit89]

Just got the shipping notification for my 2600+!  

[Dbug]

34 minutes ago, protomank said:

Hum, I have no idea on how the Stella code is structured, but it seems this could be improved, as CPU, sound and TIA can be on their own thread, even the translation from 6507 to ARM/x86 in one, to allow multiprocessing to take advantage of multiple cores.

 

Anyway, I did not notice anything like that on my PSX Classic, that is where I do run emulators, but I will take a closer look. It is a old SoC, but the best from those mini-consoles, I do not see CPU/GPU age being, per se, the issue. (picking an old and weak, of course is).

Technically, I don't think any PSX game was really timing critical, probably don't require a cycle accurate emulation of the system, plus MIPS is relatively easy to handle in term of instruction time (it's a RISC after all).
The VCS on the other hand HAS to be cycle accurate, specially with the crazy register system where the value does not matter but the "when" it was written to did.

[Thomas Jentzsch]

42 minutes ago, protomank said:

Hum, I have no idea on how the Stella code is structured, but it seems this could be improved, as CPU, sound and TIA can be on their own thread, even the translation from 6507 to ARM/x86 in one, to allow multiprocessing to take advantage of multiple cores.

This doesn't work for the 2600. Everything is tightly coupled on pixel-cycle level. We discussed this intensively in the past (when the R77 showed up). In the end, only the TV effects can be effectively spread on multiple cores.

[Brad_from_the_80s]

Probably as expected, but here and elsewhere online it seems a lot of people are finally getting shipping notices now in time for holidays, hopefully every preorder that has been waiting so long is getting theirs soon.

[protomank]

15 minutes ago, Dbug said:

Technically, I don't think any PSX game was really timing critical, probably don't require a cycle accurate emulation of the system, plus MIPS is relatively easy to handle in term of instruction time (it's a RISC after all).
The VCS on the other hand HAS to be cycle accurate, specially with the crazy register system where the value does not matter but the "when" it was written to did.

Every emulated system should be cycle-accurate, but of course, it isn't. But I was speaking not about the PSX emulation, but Stella running on the PSX Mini using Project Eris

[protomank]

I just checked HERO running in Stella (2014 core, I do not know why it has this date on it) on Emulation Station in PSX Classic through Project Eris and the movement is just fine. So, an old ARM SoC is capable of running it without that issue.

Then, I was re-watching the video, and it looks like it is jumping frames, like to compensate running on a different frequerncy than the video output. It occured me, is that a PAL running in 60Hz or vice versa, by any chance? Because that is a common result of, for example, a system that runs at 58Hz running on a 60Hz display; it runs faster but now and then, to adjust the game speed to the display frequency, it needs to drop the frame, otherwise the game would run faster then it should.

Edited December 5, 2023 by protomank

[Rodney Hester]

27 minutes ago, Dbug said:

Technically, I don't think any PSX game was really timing critical, probably don't require a cycle accurate emulation of the system, plus MIPS is relatively easy to handle in term of instruction time (it's a RISC after all).

Just being pedantic for anyone running across this in the future..._most_ PSX games aren't super timing-critical, but Gran Turismo 1 and 2 absolutely are (ironically, for replay purposes).

[Thomas Jentzsch]

6 minutes ago, protomank said:

I just checked HERO running in Stella (2014 core, I do not know why it has this date on it) on Emulation Station in PSX Classic through Project Eris and the movement is just fine. So, an old ARM SoC is capable of running it without that issue.

That core is waaay outdated. Since then the emulation core has been changed completely.

6 minutes ago, protomank said:

Then, I was re-watching the video, and it looks like it is jumping frames, like to compensate running on a different frequerncy than the video output. It occured me, is that a PAL running in 60Hz or vice versa, by any chance? Because that is a common result of, for example, a system that runs at 58Hz running on a 60Hz display; it runs faster but now and then, to adjust the game speed to the display frequency, it needs to drop the frame, otherwise the game would run faster then it should.

I would understand if it drops a frame and then jumps by 2 frames. But it sometimes drops a frame, displays the dropped frame and then jumps by 2 frames.

[protomank]

1 minute ago, Thomas Jentzsch said:

That core is waaay outdated. Since then the emulation core has been changed completely.

I would understand if it drops a frame and then jumps by 2 frames. But it sometimes drops a frame, displays the dropped frame and then jumps by 2 frames.

Yes, the core build itself is from 2020. But if Stella was able to handle the game then, it should be now. I will look to check if there is someone that made a more recent core port to ARMv8 to run it on PSX Classic.

[Brad_from_the_80s]

FWIW I have an NTSC Hero coming my way in a few days, and I know my setup is in Game mode without any additional processing, so I can (eventually) check if I can see any real issue here.  I tend to be suspicious of these reports not knowing what type of cart is in play or whether they adjusted their display properly.

[Thomas Jentzsch]

3 minutes ago, protomank said:

Yes, the core build itself is from 2020. But if Stella was able to handle the game then, it should be now.

Which game are we talking about? And handle a game doesn't mean it emulates it correctly. Today's Stella is a completely different beast.

[protomank]

1 minute ago, Brad_from_the_80s said:

FWIW I have an NTSC Hero coming my way in a few days, and I know my setup is in Game mode without any additional processing, so I can (eventually) check if I can see any real issue here.  I tend to be suspicious of these reports not knowing what type of cart is in play or whether they adjusted their display properly.

Perfect. That should settle the score.

Thank you.

[+remowilliams]

6 minutes ago, Brad_from_the_80s said:

FWIW I have an NTSC Hero coming my way in a few days, and I know my setup is in Game mode without any additional processing, so I can (eventually) check if I can see any real issue here.  I tend to be suspicious of these reports not knowing what type of cart is in play or whether they adjusted their display properly.

 

The cart is an NTSC Activision Hero cart.  Again, it's not a display issue, that is a direct HDMI capture - no display settings involved.

[protomank]

1 minute ago, remowilliams said:

 

The cart is an NTSC Activision Hero cart.  Again, it's not a display issue, that is a direct HDMI capture - no display settings involved.

Thanks for the clarification. Let's see if anyone can reproduce the issue too.

[Brad_from_the_80s]

3 minutes ago, remowilliams said:

 

The cart is an NTSC Activision Hero cart.  Again, it's not a display issue, that is a direct HDMI capture - no display settings involved.

Got it, I assume you are the source.  For whatever reason I was thinking this was in reference to some YT reviewer showing something on their TV.

[tradyblix]

Anecdotally, I am not having any serious performance issues in any of the 30 or so games i've been playing on the 2600+. The things that I have noticed is crackly sound effects in a few games and paddle jerkyness on super breakout (video olympics seemed fine tho). That's the only thing that's "bothered" me so far.

 

I don't have HERO so haven't tried it. But it is on my list of games to get. 

Edited December 5, 2023 by tradyblix

[Ben from Plaion]

7 hours ago, protomank said:

You can call me crazy, but I got more excited for the PCB than the VIP Box for a moment. The nrd engineer inside me speaks louds sometimes

Yeah its a really early 2600+ prototype PCB - it has only a 2600 socket (not the 7800 socket found in the retail release). Then it has pins and sockets for the connection between Metal Lever PCBs and the IC board rather then the ribbon cables in the retail release. I have personal attachment to the board cause it was the first working board I tested games on and I remeber that it was a great feeling to actually have something that actually played original cartridges - it was like a "shit this is really going to work" moment.

 

[DEANJIMMY]

32 minutes ago, tradyblix said:

Anecdotally, I am not having any serious performance issues in any of the 30 or so games i've been playing on the 2600+. The things that I have noticed is crackly sound effects in a few games and paddle jerkyness on super breakout (video olympics seemed fine tho). That's the only thing that's "bothered" me so far.

 

I don't have HERO so haven't tried it. But it is on my list of games to get. 

Same here, my PAL H.E.R.O. runs fine and I did not notice something critical in the same amount of games. Not talking about 7800 games.

But yes, Kaboom PAL plays a bit jerky and Super Breakout PAL even more. The nervous paddle looks like it jumps 2 pixels.

Video Olympics and Warlords run very smooth with same setup.

[John Stamos Mullet]

I get the same result with my NTSC HERO - jerky movements on and off. Not so much that it isn't playable, but definitely noticeable/distracting.