Our New Machine, Pandora

[zaoin]

The original plan was to make a low cost machine with a good spec (hires screen, lots of ram, mid range but low cost soc etc) however the view of all the threads seems to point that most people want something quite high spec and don't mind paying more for it.

Me not.
I prefer versatility to power because I see this new handeld as a new tapwave zodiac.
I would buy NDS if it had pda software.

Edited:
Added:

200€ it's my limit.

 

[A_SN]

Do we really need such a powerful device? We mainly want to play retro games so we don't need 100 million polygons for that. Also, homebrewers aren't really capable of making use of all that power as they seem to make things like Breakout clones etc. Only the really big companies have the resources to utilise that kind of power.

Homebrewers aren't capable to make use of that power? Maybe, but if we never get our hands on such powerful machines, we'll never ever make use of that much power. If I could dev on such machine, best believe I'd make use of the power, and not only for the graphics, but also for the physics.

Besides, you forgot something capital : ZOMFG N64 EMLUNATOIN!!!11111

Also, you'll only have too much power when you'll have a fullspeed XBox 360, with transparencies, but by then you might want a Playstation 5 emulator instead.

On a side note, if anyone cares, from following the on-going SoC talk (which is by the way very constructive and pertinent), I too am convinced that the i.MX31 is the best choice, and I'm glad to see craigix state that they pay a lot of attention to our comments and thus will most likely settle for the i.MX31

 

[Firefox]

I'm pulling for the i.MX31 too because there's very detailed documentation available for it that you can just download from Freescale's website.

MagicEye's shocking MMSP2 documentation, and the fact that after several weeks trying to get a copy from them I had to download it from what was practically a warez site, really put me off developing for the GP2X back in 2005.

 

[Eolair]

I'm currently undecided between the i.MX31 and the OMAP3430. The big concern for me is performance. I'm worried the OMAP3430 is too much power and cost for our needs (price comparison needed). However, at the same time I'm also worried that the i.MX31 is too little power too late. It might lack the punch needed to impress and perform for the future.

The fact is that the i.MX31 is only using the Lite version of the PowerVR MBX 3D accelerator. If you just look at the numbers it seems comparable in performance to the MMSP2+. Both are said to push 1M polygons/sec and 100Mpix/sec.

I was never impressed by the MMSP2+. Frankly it felt medicore and it didn't seem to offer significant improvement over MMSP2. The reason I prefer the i.MX31 over the MMSP2+ is the 532MHz ARM11 processor and better power management. The question is if it brings enough power? I think the answer depends on what kind of machine you want.

 

[TyBO!]

Well, since I now know that the SoC for the device is still up in the air, I'll say that my number one SoC choice would have to be the OMAP3430. If that's not a realistic possibility though, I'd like the i.mx31.

My primary concern is the life cycle of the device. Everyone seems happy with the price point of the GP2X, but now people are saying it's on it's last legs, when the GP2X hasn't even celebrated it's second birthday. That's like spending $100 a year for a gaming device.

I'm sure there are GP2X enthusiasts out there who will argue that the GP2X still has many happy years ahead of it, but lets be honest with ourselves: it will lose a lot of it's fanbase once craig's new device goes to market.

I wouldn't mind spending upwards of $300 if it's something I'll still be playing with 5+ years down the line.

Ya know?

 

[Orkie]

That's like spending $100 a year for a gaming device.

I know for a fact many people here spend more than that on consoles (myself included) .

 

[Squidge]

QUOTE

My primary concern is the life cycle of the device. Everyone seems happy with the price point of the GP2X, but now people are saying it's on it's last legs, when the GP2X hasn't even celebrated it's second birthday. That's like spending $100 a year for a gaming device.



Something is always on it's last legs when people run out of things to emulate or start hitting walls with there development. Lets face it, it wasn't exactly "leading edge" when developers were using it 2 years ago - some GP32s could equal the speed of the gp2x's stock performance.

I'm currently undecided between the i.MX31 and the OMAP3430. The big concern for me is performance. I'm worried the OMAP3430 is too much power and cost for our needs (price comparison needed). However, at the same time I'm also worried that the i.MX31 is too little power too late. It might lack the punch needed to impress and perform for the future.

For emulator development, sure, but I doub't it for homebrew games. Given the advancements in technology however, it will probably be outdated in 18 months and be unable to emulate newer systems sufficiently, but thats the price we have to pay. It takes time for mains-powered desktop systems to be emulatable by a handheld running on batteries.

QUOTE

The fact is that the i.MX31 is only using the Lite version of the PowerVR MBX 3D accelerator. If you just look at the numbers it seems comparable in performance to the MMSP2+. Both are said to push 1M polygons/sec and 100Mpix/sec.



I don't see a problem there unless people try and start emulating the PS2 and Dreamcast

QUOTE

I was never impressed by the MMSP2+. Frankly it felt medicore and it didn't seem to offer significant improvement over MMSP2. The reason I prefer the i.MX31 over the MMSP2+ is the 532MHz ARM11 processor and better power management. The question is if it brings enough power? I think the answer depends on what kind of machine you want.



The MX31 has not just got a faster clock speed, it also has far more cache (160KB compared to 32KB), on-die sram (16KB), a full 32-bit ddr memory interface (compared to the 16-bit interface on the mmsp2+), etc.

 

[Tobriand]

Basically, it depends on how future-proof the device is to end up. If we're after something that does everything we wanted the 2x to be in the first place (and not much more) then the iMX31 should be all we need, if I were to guess right. I'd guess - from a non-coder point of view - that it would allow us ample power to get at full-speed (or nigh on) Amiga and PSX, let alone eeverything that's there or nearly there already.

If, on the other hand, we next want to be having this conversation in 5 years' time as opposed to 3, then the OMAP chipset might well be the one to go for, and just swallow the price.

I suppose the other issue to consider is this: When do we think mobile chipsets capable of emulating current/last gen platforms is likely to appear? I mean, OK, PS2/Xbox aren't exactly amazingly high on the list off things we need in a handheld, but from the emulation point of view, they're probably what the handheld after this one might be aiming at. Which means by the time it's developed, there has to be a chipset capable of dealing with those systems. And is that likely to have haappened in 3 years? Personally, I think not... Maybe not even in 5 or 10 come to that...

Just a thought...

 

[Eolair]

Regarding the concern about battery life.

I think it is likely that the OMAP3430 has lower or at least equal power consumption compared to the i.MX31. One reason being that it uses a Cortex-A8 core and a 65nm process, another reason being that it takes advantage of TI's SmartReflex (dynamic voltage, frequency and power control). So I don't think we have to worry about that point. Both SoCs should be much better than the MMSP2+

 

[Eolair]

Right now I think it very much depends on the price difference and how much the OMAP3430 would change the end price of the device. If it doesn't significantly increase it I would say go with the OMAP3430, if it does then you might want to consider if you really need that extra power.

 

[sam fisher]

That's like spending $100 a year for a gaming device.

I know for a fact many people here spend more than that on consoles (myself included) .

I concur. I would happily pay a large price for a better device upfront

 

[Javacat]

If the OMAP3430 really is as powerful as a few people are making out, but does push the price over a certain point then it could be the perfect excuse to go and build something incredible (whilst putting out of a few peoples budgets) :ninja:

 

[sam fisher]

If the OMAP3430 really is as powerful as a few people are making out, but does push the price over a certain point then it could be the perfect excuse to go and build something incredible (whilst putting out of a few peoples budgets) :ninja:

Indeed h34r:

 

[X-Code]

Amiga was expensive and incredible... Give me a handheld that "feels" the same and I'll buy it

 

[jadon]

I want this community to make an OMAP3-based device more than anyone and I think it would be awesome. If it can't be done at a reasonable price, then I think it'd be right to forget it. I can't give out pricing, but I can work with those who can. As soon as Craig and I can manage to talk, we might be able to figure something out.

Amiga was expensive and incredible... Give me a handheld that "feels" the same and I'll buy it

 

[Exophase]

The fact is that the i.MX31 is only using the Lite version of the PowerVR MBX 3D accelerator. If you just look at the numbers it seems comparable in performance to the MMSP2+. Both are said to push 1M polygons/sec and 100Mpix/sec.

I think those specs will give you a lot more mileage on i.MX31 than MMSP2+ because of the tiled rendering architecture.

 

[God Ginrai]

If craig was to choose the OMAP, I'd easily pay $350 for the handheld. (possibly even more) I don't really care about emulating the PS2 or XBox or Gamecube, but N64 emulation capability would be a definite seller to me. There's nothing I want more than handheld Ocarina of Time. (Not to mention, all the other games in the N64 library that would be playable) Plus, all the PSX fans would get themselves a fullspeed PSX emulator most likely.

I understand the argument about how much power homebrew games use, but what about all the open source Linux games out there that we could try to port? I'm sure there are a lot that people really like that would use a decent amount of processor power.

Not to mention, the OMAP says it handles H264 with ease. Nice quality for your video, no?

Plus, with more processor power, you could have more programs running in the background so you could multitask easier.

-God Ginrai

 

[argor]

i have to say the i.MX31 is more wise chose

 

[Rivroner]

If the final cost between i.mx31 and Omap3430 was about + 40 $ , do you really want a worse soc? I think that if the final cost is not more than 50 $ the choosen one have to be Omap.

If you are gonna spend 220 $ you can easily spend 260 $ for a better and future inversion.

I vote for Omap3430, but i would be happy with i.mx31 as well, i like it; but we always want to have a bigger thing

 

[God Ginrai]

i have to say the i.MX31 is more wise chose

It really depends to me, but I definately think that the MMSP2+ shouldn't even be considered any more, that it should be between the i.MX31 and the OMAP. My vote would be for the OMAP, but I don't think that choosing the i.MX31 would be a condemnable action like some people who want more and more power might say. I just really think that geting fullspeed PSX emulation and N64 emulation is worth the extra bucks for the OMAP.

If the final cost between i.mx31 and Omap3430 was about + 40 $ , do you really want a worse soc? I think that if the final cost is not more than 50 $ the choosen one have to be Omap.

If you are gonna spend 220 $ you can easily spend 260 $ for a better and future inversion.

I totally agree.

-God Ginrai