Graphics: Turbo vs. Genesis - ye old debate

[Black Tiger]

That's a fun dream, whereas the Turbo REALLY IS an 8-bit console.  Last time I heard, the X-box 360 was getting crushed by the lame-assed Wii.  No I am not this nintega person.  I would never disgrace myself by using a name that has any part of the word nintendo in it.

The Intellivision REALLY IS a 16-bit console by the same standards.

[spenoza]

I like to think of my Xbox 360 as an 8-bit console, because it crushes the puny competition even more so.

That's a fun dream, whereas the Turbo REALLY IS an 8-bit console.  Last time I heard, the X-box 360 was getting crushed by the lame-assed Wii.  No I am not this nintega person.  I would never disgrace myself by using a name that has any part of the word nintendo in it.

How do you define bit-ness? The CPU is 8-bit, but the graphics chip is fully 16-bit. Further, the CPU in the Genesis is 16-bit, but I suspect (maybe Tom can clarify) that the data bus is only 8-bit. And really, bit-ness doesn't mean much. The SNES has a 16-bit CPU, but the core design is still very similar to the 8-bit CPU on which it is based. And when you take an objective (well, in a relative sense, anyway) measure of CPU performance, like MIPS or somesuch, you find that the CPUs in the Genesis, SNES, and TG-16 all perform on a similar level. So is an 8-bit CPU that's as "powerful" as a 16-bit CPU a lesser CPU because it is 8-bit? Or is it a peer CPU because it can chug out just as much real-world performance?

See, the whole 8-bit console vs 16-bit console thing is really more about perceptions than technical details. It's about lumping together games of similar style and make rather than technical specifications, ultimately. There's no argument that some of the very early TG-16 games were largely 8-bit in scale and scope and meant clearly to compete with the NES. But the later games on the system were very clearly the peers of any of the titles on the SNES or Genesis. It's hard to argue that games like Art of Fighting, Magical Chase, Aero Blasters, Lords of Thunder, etc... are somehow of the same generation of development, graphics, and audio as the NES. These games are clearly peers to anything released on the "true 16-bit" systems. So in this sense, the PCE is a bridge system. It's the system that connects the 8-bit generation to the 16-bit generation. Just as the Neo Geo bridges the 16-bit generation with the 2D games of the 32-bit generation. The N64, with its 64-bit CPU, is clearly of the 32-bit generation regardless of its technical details, whereas the Dreamcast is considered, at least timing-wise, the bridge system between the 32-bit generation and the following generation (PS2, GC, Xbox).

I know I'm beating a dead horse here, rehashing stuff we all largely agree on. There's no point arguing the bit-ness thing any further. Regardless of particular design decision trade-offs and the lack or presence of a particular capability, the PCE is definitely a peer system to the Genesis and SNES. On this I think there is no meaningful debate. It doesn't matter if any one of us likes one system a little better than the others, they're all clearly in the same ballpark and playing the same game, if you will.

[ccovell]

The 68000 is actually a 32-bit processor, at least as far as internal registers and [macro]instructions are concerned.  However, over its lifetime, different varieties of the 68000 had data buses ranging from 8 bits wide to 32 bits.

[Black Tiger]

I think that "amazing for only 8-bit!" comments are more appropriate for SMS games, many of which feature graphics and sometimes sound equal to or better than many 16-bit gen console games.

There's no argument that some of the very early TG-16 games were largely 8-bit in scale and scope and meant clearly to compete with the NES.

There are many Gen/MD & SNES/SFC games with 8-bit-like aspects, which aren't all early efforts either.

[spenoza]

The 68000 is actually a 32-bit processor, at least as far as internal registers and [macro]instructions are concerned.  However, over its lifetime, different varieties of the 68000 had data buses ranging from 8 bits wide to 32 bits.

I tried to assert something like this some time back and Tom took me to task for over-simplifying. But yes. The 68000 as used in the Genesis and early PC varieties was tied to a 16-bit bus, I believe. Just goes to show how little all this bit-shit actually adds up to...

[spenoza]

I think that "amazing for only 8-bit!" comments are more appropriate for SMS games, many of which feature graphics and sometimes sound equal to or better than many 16-bit gen console games.

There's no argument that some of the very early TG-16 games were largely 8-bit in scale and scope and meant clearly to compete with the NES.

There are many Gen/MD & SNES/SFC games with 8-bit-like aspects, which aren't all early efforts either.

This is very true. From my perspective, however, there are a greater proportion of games (among chip titles, at least) on the TG-16 which exhibit those qualities. It has less to do with the technical qualities of the system and more to do with the timing of its introduction and the conditions of the market at the time, I suppose.

[nodtveidt]

The whole bit war thing was stupid from the start and had no bearing on how the games were. I've played many an NES game that was superior to anything I've seen on the PS3. The bit thing was just marketing hype that they eventually had to abandon once consoles got too complex to measure in bits.

[Tom]

The 68000 is actually a 32-bit processor, at least as far as internal registers and [macro]instructions are concerned. 

  Even Sega and SNK didn't hype that in marketing Paired 16bit regs and a 16bit ALU, never mind the 16bit data bus.

 Funny thing, a standard random access memory read, add, store of 32bit element is slower on the 68k(even with its macro instructions) than on the 65816 and that's a 16bit processor on an 8bit bus. And that's with the 65816 using *no* DP/ZP regs. Sequential free incrementing on the 68k though brings the cycle counts ,between the two, even - which is still surprising for the 65816 considering it has no macro instructions or pair 16bit regs or auto-increment. Some people also boast about it having the ability of moving 32bits at a time for source/destination, yet fail to see the 65816 and 6280 block move instructions are faster at the same clock speed. Among other examples.

 I wrote a 68k hsync routine for the PCE. The idea was to drop in a 68k in place of the 6280, to see some direct comparisons. Surprisingly, the 68k code ended up being 49cycles longer or 35% percent slower (93 vs 142). And that was with reserved half the registers of the 68k - which is going to cripple it fairly good for other game code because it's more of a register-register processor than the 65x arch. Realistically, I wouldn't reserve that many registers and the cycle count would go up to about 160+ or more. If that was a full screen sine or such, you'd really start to eat into cpu resource.  And my 6280 code could have been 80 something cycles for some instances(if you only need to change the lower 8bit part of the scroll reg) or even 70 something cycles for arcade card optimization. That's definitely not going to be the case for every situation. But that's a good example how one can optimize for 8bit elements to drive that speed.

The whole bit war thing was stupid from the start

 Yeah. And it still goes on.

[Blammo]

The whole bit war thing was stupid from the start and had no bearing on how the games were. I've played many an NES game that was superior to anything I've seen on the PS3. The bit thing was just marketing hype that they eventually had to abandon once consoles got too complex to measure in bits.

It might have been bogus but it wasn't stupid - at least on Sega's part anyway, and on Atari and Commodore's in the computer market. It was a good marketing move that summed up in a single term what made - or was alleged to make - the Genesis better than the competition.

Perhaps NEC shouldn't have pushed it quite so much though. I do wonder how much of it was an attempt to compete with Sega though: by the time NEC started shouting about 16-bit with the Turbo's release in '89, the Japanese Mega Drive with its massive 16-Bit sign had already been out a year. I don't remember ever seeing "16-Bit" being used in relation to the PC Engine. It certainly seems as if pushing the 16-bit angle didn't occur to them until Sega started it.

[Joe Redifer]

Putting "16" in the name of the US console was NEC's way of marketing against Sega's 16-Bit campaign.  They didn't want to be seen as "less bits" and therefore less powerful  Marketing can be a powerful tool.

[Gentlegamer]

Putting "16" in the name of the US console was NEC's way of marketing against Sega's 16-Bit campaign.  They didn't want to be seen as "less bits" and therefore less powerful  Marketing can be a powerful tool.

Now explain the "TurboGrafx" part.

[Necromancer]

Now explain the "TurboGrafx" part.

Turbo = more power and Grafx = graphics, so TurboGrafx must = more powerful graphics.  What about that isn't obvious?  It's not like they named it something completely groundless (such as Xbox or Wii).

[Tom]

Putting "16" in the name of the US console was NEC's way of marketing against Sega's 16-Bit campaign.  They didn't want to be seen as "less bits" and therefore less powerful  Marketing can be a powerful tool.

Now explain the "TurboGrafx" part.

 I think that decision came from Japan. I mean, SuperGrafx and CoreGrafx names came out the same time frame as TurboGrafx.

[Joe Redifer]

I always thought that TurboGrafx-16 literally translated into "Fast 16-Bit Graphics!!!!"  Very descriptive and actually not really deceptive at all.

[guyjin]

Actually, I think the TurboGrafx-16 has the best name of the 16 bit consoles: "Super Nintendo entertainment system", while it gets the point across, is not terribly interesting. And "Genesis" doesn't even make sense. Did they want us to think of the Bible or early 80s british bands? what do those things have to do with videogames? "mega drive" was better, but not as cool as "TurboGrafx-16".