MSX Turbo-R : Unexpressed potential?

Page 1/3
| 2 | 3

By journey

Hero (537)

journey's picture

23-10-2021, 10:18

In your opinion, to date, the potential of the MSX Turbo-R has been fully exploited?
What is the game that makes the best use of the machine?
For example, these days I'm playing Seed Of Dragon, and I just don't understand how it's possible that a Turbo-R specific game doesn't use hardware scrolling (same for Illusion City and Fray).

What do you think?What do you think?

Login or register to post comments

By ARTRAG

Enlighted (6584)

ARTRAG's picture

23-10-2021, 11:21

The Turbo-R has for sure a lot of unexpressed potential but it is also affected by terrible crappy decisions in its HW design. The I/O speed toward the VDP is the most critical flaw, where in Z80 mode you get an higher bandwidth (!). This means that you have a very fast CPU waiting the most part of the time to the VDP, unless you do acrobatic coding to interleave other tasks with I/O... Also it lacks of programmable interrupts and this makes the PCM audio just a useless frill. The result is that games using it will hang during playback (as the first experiments using the key click to play 1 bit audio) unless you do acrobatic coding to interleave audio with other tasks.

In the end, the potential is well hidden and hardly accessible. Without coding strategies more suitable for demos than to games, you cannot go very far from an MSX2+

By Manuel

Ascended (18256)

Manuel's picture

23-10-2021, 12:04

Maybe Moon Light Saga is a better example of good use?

By journey

Hero (537)

journey's picture

23-10-2021, 12:45

Manuel wrote:

Maybe Moon Light Saga is a better example of good use?

For the scrolling for sure.
But I have never seen a title that made me escalate
"Well, this could never have been done on MSX2 / 2 +"

By gdx

Enlighted (4833)

gdx's picture

23-10-2021, 13:54

The Turbo-R allows us to do more things mainly thanks to its speed and its larger internal memory, but if this has not been exploited, it is because of the lack of enthusiasm of the developers.

No relation to the subject but ...

ARTRAG wrote:

The I/O speed toward the VDP is the most critical flaw, where in Z80 mode you get an higher bandwidth (!). This means that you have a very fast CPU waiting the most part of the time to the VDP, unless you do acrobatic coding to interleave other tasks with I/O...

It is mainly the VDP and especially its commands that is slow.

ARTRAG wrote:

Also it lacks of programmable interrupts and this makes the PCM audio just a useless frill.

What PCM lacks is rather a dedicated memory like that of MSX-Audio. It is incomprehensible to have chosen this PCM instead of MSX-Audio or compatible with more PCM tracks. Fray's sound digitizations are a pity for what we can do with MSX-Audio.

By Sandy Brand

Master (245)

Sandy Brand's picture

23-10-2021, 15:38

gdx wrote:

The Turbo-R allows us to do more things mainly thanks to its speed and its larger internal memory, but if this has not been exploited, it is because of the lack of enthusiasm of the developers.

No, as ARTAG was trying to point out is that the Turbo-R architecture goes against sane hardware design: faster CPU won't matter if it is stuck waiting on a lot of stuff. I would even argue that not having a VDP interrupt on command completion in the V9938 was already a major flaw: it means your architecture will struggle heavily to achieve full parallelism between CPU and VDP. Blaming it on the developers is easy but rather unfair Smile

gdx wrote:

It is mainly the VDP and especially its commands that is slow.

The command speed is rather irrelevant here. What the system needed was a better VDP with better character modes that are more tailored to games. If you look at console hardware around that time you can clearly see the trends there and for good reasons. Instead of issuing complex VDP commands that it needs to wait on for completion, the CPU then only needs to manipulate character IDs in video memory, which is much faster and less convoluted to code.

A partial example for reference: the SNES CPU speed is comparable to that of a Z80. Smile

By NYYRIKKI

Enlighted (5889)

NYYRIKKI's picture

23-10-2021, 16:43

Does MSX tR still have unexploited, hidden potential? For sure it has!

One of the interesting details include the DMA and related functionality. Unfortunately those have been disabled in hardware by tying the CPU pins to ground and therefore making these features inaccessible to programmers or less intrusive hardware developers. The CPU is also quite a bit more capable to do 16bit calculation compared to Z80, but indeed the output is the real bottle neck. I do agree that V9958 is just too slow to talk "in real time" with CPU. You can actually do a multiplication in between two I/O operations "for free" as the CPU would stop anyway to wait VDP... Can you actually use this "free time" to anything useful is a question that I think still waits for an answer. There are even more, less important things to cover, but I think these are the things that could make at least somewhat noticeable differences if they would be exploited properly. I think the V9958 is a bit "unexploited territory" by it self. We all know it can do smooth side scroll more easily compared to V9938, but are we really sure there is not more we should know?

By gdx

Enlighted (4833)

gdx's picture

23-10-2021, 16:51

Sandy Brand wrote:

No, as ARTAG was trying to point out is that the Turbo-R architecture goes against sane hardware design: faster CPU won't matter if it is stuck waiting on a lot of stuff.

I did not say that. I talked about the VDP. I said that the biggest problem is its slowness, not interruptions.

Sandy Brand wrote:

Blaming it on the developers is easy but rather unfair

I don't blame the programmers. I just note that programmers have always preferred to develop for MSX1 and for MSX2 to a lesser extent. This is normal given that the 2+ and turbo R are much less common or even almost non-existent in many countries.

Sandy Brand wrote:

The command speed is rather irrelevant here. What the system needed was a better VDP with better character modes that are more tailored to games. If you look at console hardware around that time you can clearly see the trends there and for good reasons. Instead of issuing complex VDP commands that it needs to wait on for completion, the CPU then only needs to manipulate character IDs in video memory, which is much faster and less convoluted to code.

You are wrong, V9938/58 commands are not complex, just slow. Fast copies (or at least tiles) and better sprites are essential when the CPU is a slow 8bit to make games.

Anyway, there are obviously plenty of things that could have been better, but that doesn't answer Journey's question.

By Manuel

Ascended (18256)

Manuel's picture

23-10-2021, 16:49

Does V9990 help to get more potential out of the turboR?

By gdx

Enlighted (4833)

gdx's picture

23-10-2021, 16:51

Yes, sure.

By geijoenr

Champion (276)

geijoenr's picture

23-10-2021, 16:52

I was thinking that indeed a VDP on a cartridge with superimpose and a direct VRAM access from the CPU would probably made a turboR a way more capable machine.

Page 1/3
| 2 | 3