Nuon, the gaming chip that nearly changed the world but didn’t | Hacker News Reader

Nuon, the gaming chip that nearly changed the world but didn’t

arstechnica.com

86 pointslermontov29 comments

I've been following it and even done some development around 98-99. It was a bizarre system with tools that were not at the level even with the 5th/6th gen consoles available at the time. You pretty much had just a gcc port with some serious bugs, an optimizing assembler (it's VLIW so the assembler had -O option and the bigger number you put there, the better it will optimize, at 6 though it was running out of memory on any non-trivial code) and a broken debugger, written in XLisp for some reason, which I never managed to use productively.

The thing had 4 PUs, each with local memory and a DMA controller to access system and video memory. You could run one of them in a CPU emulation mode, where local memory was used as cache and it was the only target the C compiler supported. If you wanted local memory and/or more than 1 PUs - you wrote assembly and debugged by writing into the frame buffer.

Even then I was wondering how it's going to compete with the DC/PS2 it was coming up against but my bosses told me it's fine because it's going to be in every DVD player and everybody is going to by a DVD player pretty soon so the game consoles are pretty much already dead.

Thanks for the reality check :-)

It's really funny, because all throughout reading that article I was thinking "this thing never had a chance". A custom architecture based on principles that are notoriously difficult to write tools for (VLIW), probably adding 50% to the BOM of very price sensitive consumer electronics, and an attempt to create a gaming platform by a company with no capability of producing first party titles and no commitments from credible publishers to do third party games either. How could that not fail?

But then the article ends with all of these people believing - even with 15 years of hindsight - that they genuinely were just on the verge of breaking through. Their technology was just too far ahead of its time and people didn't yet understand it!

To be fair, the game consoles were not a walk in the park either. PS2 VU programming was in some senses more complicated - on NUON you just had to stuff as many slots in a packet as possible e.g. you could write a jump to self and had like 20 slots for instructions in there (the jump packet itself and 2 delay packets, each was 6 or 7 instructions possible if memory serves) but on PS2 you had to be more clever with hiding latency between dual pipes. A loop on PS2 would be processing several items at once, eg stage 1 for the current idex, stage 2 for the previous index and the final stage for the current index - 2.

Programmers at the time were okay with this and it was actually fun. The tools though were much better for the competition, on PS2 I could set break points anywhere (even on memory reads) and see the DMA streams in the debugger, on NUON it was pretty much debugging with prints except you had to write the print yourself :)

Game developers go where the market is, regardless of the technical challenges, if they got their chips shipped with enough dvd players, people would start making games for them and it would snowball. Would it replace dedicated gaming hardware? Probably not, but it certainly could of eaten some of the casual market like the iphone did to the gameboy.

Yes, non-gaming platforms will eventually get games just by being popular. But that requires millions of units at a minimum, and even that might not be enough (what was the state of desktop Linux gaming in 2012?). For this path to work, the platform needs to become popular first. For the PC the early value proposition was productivity applications. For the iPhone it was good mobile web browsing, maps, etc.

A gaming system (which this was) can't wait for that popularity to happen through some other means. The games are the raison d'être of the system, the only reason for people to buy one. What was the end-user pitch here that'd get one of these to the hands of millions of consumers? "You can buy some other DVD player for $200, or this one that might some day play games for $300". Not very compelling, when both of them do an equally good job with the main task you're buying a DVD player for.

> Yes, non-gaming platforms will eventually get games just by being popular.

I wouldn't say this is an inevitability. Look at Nokia feature-phones.

Sure, from what I've heard and read an incredibly poorly designed ecosystem and lack of carrier interest, but you can't tell me Snake was the best the market could come up with...

>>How could that not fail?

The dev team got paid for ~7 years?

>Miller had been heavily involved in the creation of Atari's 64-bit Jaguar game console, which was already petering out after just a year on the market. "We were very proud of it," he said in a recent interview. "But it did not succeed because of the lack of content and a chicken and egg issue

noo, Jaguar was TERRIBLE to program, and from what pandaman writes Nuon was just the same.

> "I kind of hoped that it would be a real computing platform, much as the Atari ST had been a platform that was used for games as well,"

Atari ST was also pretty TERRIBLE, it was used for games _despite_ being terrible, not because it was a good computer.

>He also brought in lawyer

and this explains why it failed. 3DFX history was similar, they also toured Japan, but instead of lawyers they got a finance guy, they build hardware themselves, none of that 3DO CD-I licensing rubbish lawyers love so much.

From what I've heard from people who have seen both, hardware was similar to the Jaguar too.

This is fascinating to me. It's worth noting this was from the same console generation as the PS2, Xbox, Dreamcast, and GameCube.

I found some specific information about it here [1]

> The technology in the few available NUON DVD players raised the price around $100 over comparable players.

> Only eight games were ever released for NUON. The graphics are comparable to early PlayStation 2 games and are most notable in Iron Soldier 3, a mech combat game.

Here's a video showing the Iron Soldier 3 graphics [2]

> The best game is generally regarded to be Tempest 3000 (many sources consider it the only truly playable game)

The wiki also lists 16 homebrew titles including an Atari 800 emulator and Doom.

On the NUON-Dome site [3] there's a homebrew release announcement from 2014, as well as Amazon links where you can still buy some discs and hardware.

It also features a "Video Light Machine" [4] which as far as I can tell is Winamp-style audio visualization.

[1] http://www.giantbomb.com/nuon/3045-85/

[2] https://youtu.be/0nypv2_qvSg?t=173

[3] http://www.nuon-dome.com/

[4] https://www.youtube.com/watch?v=4nJlY4Z8v-A

Great links. Adding one more - Tempest 3000 gameplay: https://www.youtube.com/watch?v=YyaeUNAAaXc

Seems like a 3D version of space shooters...?

it actually looks like a pretty nice graphically updated version of tempest (my favorite arcade game probably) in general, and pretty similar to Tempest 2000 (I vaguely remember buying it on PC way back when).

I wonder what kind of controls it had, might have been interesting if any DVD player had a jogwheel with which to simulate the original spinner control...

The Nuon gamepad was wired, and was similar to the N64 controller (D-pad on the left, analog joystick in the middle, various buttons on the right).

It had four completely different controllers. One like the PS1 with no analog stick, another like the N64. The Logitech aftermarket one was more like an original Xbox controller with D-pad and analog on the left and six buttons on the right.

This console had only eight games. One was region locked to South Korea (and you couldn't play the other seven games in that region). Another one of the games didn't even work properly on all of the controllers.

That looks more like PS1 graphics rather than PS2.

One of the main developers hired to write games for Nuon was actually the legendary Jeff "Yak" Minter, he of the Attack of the Mutant Camels fame. He wrote Tempest 3000 for it, as well as the Video Light Machine. Both were sequels to previous work; Minter wrote the original Tempest 2000 (itself a sequel to Tempest), and the VLM for Nuon was version 2.0 of a synth he did in the 90s.

I wanted one just for the VLM, to replace the one from my Atari Jaguar :)

3DO came out with a multimedia machine that played music CDs, photo CDs, could play video CDs with a daughter board. The original 3DO player was released in 1993 making it the first non cartridge video game player by several years.

The M2 was being designed and prototypes built in 1995-1996. M2 featured a dual core powerpc and was designed to ship with a DVD player making it truly the next generation player. The plans to the device were sold to Matsushita, they demoed it at E3 in 1996 but failed to bring it to market.

3DO suffered a lot of the same troubles as Nuon except about 5 years earlier. 3DO went out of the hardware business in 1996 but continued publishing some very successful and revolutionary game titles before finally going out of business completely in 2003.

Point being this has all happened before, and it will happen again.

Sounds a lot like Philip's CD-i [1] but a decade later to the party. In was in the living room with games, video (VCD) and internet access...

[1] https://en.wikipedia.org/wiki/Philips_CD-i

It had a competitor, the Commodore CDTV (basically an amiga). https://en.wikipedia.org/wiki/Commodore_CDTV Pretty cool machine, CD rom drives were sparse in 91.

Techno-optimism at its silliest. Even if it had succeeded it wouldn't have "changed the world."

Interestingly the converse (game consoles that also play DVDs) were much more successful. Not sure how much that feature was really used, and today it's pretty much moot because online streaming has supplanted physical media for most people.

The consoles want in on the online streaming market too, though. The Xbox 360 has sports streaming from MLB.tv, ESPN, etc, and Netflix support is on the major consoles.

It's interesting how fragmented from an end-user POV the streaming services are -- it's like a 100-way betamax vs VHS fight. I suppose you could argue there's only one "format": the web browser, but that's not true in the home theatre/console/smart tv space.

When it came out the PS2 was a rather cheap DVD player considering it also was a gaming console. I think the PS3 was among the cheaper if not the cheapest bluray player when it was released. And I found pretty cool that my ps1 was able to play audio cds with a neat interface.

This has always been a strength of Sony. I remember it was one of the reasons I ended up getting a PS2.

Not sure how you'd do that now given that the cost of adding Netflix or streaming support is negligible, but Sony definitely made smart moves with their timing and put forth the proposition that "For a little more than a DVD/Bluray player, I can get a PS2/3."

I think it's pretty interesting that people at Sony were upset with Kutaragi for putting DVD functionality in the PS2. People argue today that it was a small part of the mass adoption of the format. Perhaps, along with the disastrous unveiling and launch of the PS3, it contributed to his downfall at the company.

The controller fail got me smiling:

https://youtu.be/3lpViVaMtrM?t=570

I worked there in 2000 as a graphics engineer - it was an interesting tech, lots of bright people.

I wonder how this affected Toshiba's own processors -- MeP, then Cell. Some of MPEs (Nuon's processing units) only had access to its small local memory and needed DMA to access the main memory. This is kind of similar to Cell's SPEs. Each SPE has only 256 KB of directly accessible memory and required DMA to get to the main memory. You can say it's not uncommon in DSPs, but it's still interesting Toshiba has been involved with these.

Reply on news.ycombinator.com