SoC: Back and forth!

[Grench]

May be I'm too cynical, but working with Intel as such a small entity can be dangerous.

How do you figure? At no point in this process is it even necessary to directly engage Intel themselves. The data sheets are published on their website for anyone. The SoC itself is available from Mouser.

First there's the possibility of missing out on all the great things arm could have in the future,

Blanket statement that could apply to any architecture. Intel is moving from 22nm to 14nm this year.

we may need to switch back to arm at some point;

Switching from OMAP to Snapdragon to Bay Trail to etc... - it's all similar. All require redesigning the motherboard. Many of the features on the board post the SoC are similar no matter which is used.

second if Intel wants to screw us over just right after Ed has spent all his capital on making a fully functional and finalised product, it can and will.

Why would you assume they would be malevolent? Because they're huge? Intel is huge. It's like steering an aircraft carrier. They're not going to change course over this project. It takes years for them to change product plans. To 'screw us over' they would have to stop their 2 year mobile AND embedded roadmaps and buy every SoC from every reseller and screw the world. Which would be rather silly since they just released the most open mobile SoC in modern history.

We will have no place to back up to at that point because there's not as many alternatives in x86 as in arm.

At that mythical point where super-agile Intel removes all of their mobile products from the market and reclaims all items in the reseller pipeline. Mmmm...K.

It's also worth of mentioning that it is the developers in the user base made pandora different, not the end users. It's a sounder approach to try our best to keep the devs happy and give them meanings to stay.

Agreed - I like having devs interested in the project too. I'd like to make it easy for them with well documented hardware and open source drivers for everything including video. That would be awesome!

Leaving them with much to do, much to solve is better.

Hold on - did you just state the the more bugs we have and the harder they are to work on is -better- for the project and end device? Whhhaaat???

Staying with arm will at least keep the pyra unique if only as a niche product,

Niche as in elite as in new user hostile boards and device that can only be run if you read 300 pages of forums to figure out how to set it up?

jumping the hype train may convert it into another generic x86/windows smart device.

Hype train? No hype, just facts. Read up on the facts, you'll be amazed with the results.Outright fabrication: "Generic x86/windows smart device" The X86 crowd wants a Pandora style gamer's handheld on 64bit Debian on a z3770 SoC. Do you have any concept of how not-generic that really is?

I would say if Ed has millions to spend then he can do what ever he could, but with only one shot at it, staying simple and stupid is always the best approach. (Engineering 101)

And three statements earlier you wanted to make sure that it was complicated and hard to work on so that the devs had more to do and felt more loved. You want simple and stupid? Intel makes a NUC (New Unit Computing) box using a N2820 SoC which is in the same family as the z3770. Linux Kernel 3.14 installs and runs it out of the box with X and acceleration. http://www.phoronix.com/scan.php?page=article&item=intel_baytrail_nuc&num=1
How's this - make a system that, for the most part, the core system simply works - then the devs can move on to fixing and improving on all of the 'other bits' that are sure to get added onto the system, porting over orphaned PND programs that don't already exist in the Debian repository and optimizing to the new hardware. Do this by picking hardware that is already in the Linux kernel (3.14 and above), is well documented (full fact sheets), etc...

 

[Natsu]

So you're afraid that people are going to get a $700 palm top Linux OS computer confused for a full desktop with Widnows 8 and twin bleeding edge graphics cards?

correction: more like a 700 euro computer

I already start having nightmares of $700 ( ) and now you trying to kill me. :lol:

 

[xiongxioi]

I was not meant in ill-will, trust me Grench.


Intel would not stop their production line, but they can refuse to sell us any socs some time down the line. Pyra is too small a project for them to care.


I would prefer a x86 pyra if there's no major risks involved, since I can simply drop all the software in, and getting rid of my laptop.


I suppose this is a sort of expectation issue. If there was no pandora, no arm community, no gp2x; it wil be perfectly rational to make pyra x86. X86 is great, and Intel's socs are supreme in many ways. But pandora started off as the ultimate emulation handheld and it got its starting fund and all that support from the gp2x community. As long as pyra is made, no matter how much better it could become, it will be a given success. The market is already there. Even if ed ends up using cortex a9, I would still buy the pyra. There's no need to take any additional risks, even if it's very likely to be even better in the long run.


Of cause, if Ed's suddenly got a lot of support for pyra to go x86 from people who would not buy arm pyras, that will be a totally different story. I prefer pyra to be x86, but I will buy the pyra regardless. It's not about making the best product, it's about enabling ed to continue his business and keep the community alive.


Arm being cryptic and hard to use is more or less compensated by its target consumers are mostly knowledgable and educated people who are willing to learn, it's not a selling point. X86 is arguably superior on many fronts, it's ashame really. I say if pyra performed well, and thus giving ed more capital to spend on the next project, he should try it out. Ed's still owing investors a lot, he need to make some quick and safe profit before venturing into the unknown.

 

[CrazyIvan]

So you're afraid that people are going to get a $700 palm top Linux OS computer confused for a full desktop with Widnows 8 and twin bleeding edge graphics cards?

correction: more like a 700 euro computer

I already start having nightmares of $700 ( ) and now you trying to kill me. :lol:

That's why I think that people'll think twice (if not more) before ordering.

Price may decrease in a couple of years though.

Arm being cryptic and hard to use is more or less compensated by its target consumers are mostly knowledgable and educated people who are willing to learn

Just look at people that are heavily using their arm devices. Extremely intelligent people they are, like apple fans,android fans,touchwiz haters and others that have their highly technical discussions on any site or forum that somehow related to tech nowadays. Even here.

Instruction set is close to irrelevant nowadays as most people use high level languages and there will be a lot less people that need to use it in the future (the same happened in PC world in the last 15 years)

 

[Farox]

I was thinking about modularized hardware....a very nice idea.

Imagine a base with only the case (with DPAD, buttons, nubs, keyboard, external connectors, battery and maybe display ) and a daughter card with the SOC (CPU+GPU and RAM) that is changeable. At first we could start with the OMAP5 , then we could upgrade to a new ARM SOC (if we find a company that would supply to us) or another SOC (who said x86...) in few years.

The only problem will be software side...with compatibility.

 

[Klumpen]

I see another wave of upgrades coming.

At Pandora2 Pyra it's OMAP5 to Snapdragon or something similar for a couple of hundred € then?

Fragmentation dangers are growing (especcially on the software side this time). Oo

I don't necessarily want the fastest SoC, I care more about temperature and battery drain.

 

[SeeThruHead]

I could see this selling way more units if it was x86.

 

[_jr_]

Then that is your chance to get rich.

 

[hns]

I vote  that a PYRA is created with MIPS, x86 and ARM SOCs.  When it boots, it randomly chooses a SOC.

You forgot m68k, powerpc and s390. And 6502 for direct C64 emulation.

If we happen to collect enough money we can develop our own SoC that is capable of emulating all of these instruction sets by simply swapping microcode

It will be available in 2024 and cost zillions of $$$ or an orchestrated work of many volunteers. But who cares - as long as it is the best option we can get today...

 

[Askarus]

I vote  that a PYRA is created with MIPS, x86 and ARM SOCs.  When it boots, it randomly chooses a SOC.

You forgot m68k, powerpc and s390. And 6502 for direct C64 emulation.

If we happen to collect enough money we can develop our own SoC that is capable of emulating all of these instruction sets by simply swapping microcode

It will be available in 2024 and cost zillions of $$$ or an orchestrated work of many volunteers. But who cares - as long as it is the best option we can get today...

That's what we'll do when we got rich with the Pyra .

I see another wave of upgrades coming.

At Pandora2 Pyra it's OMAP5 to Snapdragon or something similar for a couple of hundred € then?

Fragmentation dangers are growing (especcially on the software side this time). Oo

I don't necessarily want the fastest SoC, I care more about temperature and battery drain.

I would not go the way to offer a Version with SoCs of different companys.

OMAP 5 to OMAP 6 might work.

But no OMAP to Snapdragon or vice versa upgrade.

That's also a reason why I would like the OMAP.

We can use further TI SoC.

 

[ekianjo]

I could see this selling way more units if it was x86.

You forget that it's a small company we are talking about, and selling a lot could lead to massive organization issues, too (and support needed).

See what happened with the Raspberry Pi ? Demand exceeded supply and people waited more than one year to get their order.

 

[wej]

Intel would not stop their production line, but they can refuse to sell us any socs some time down the line. Pyra is too small a project for them to care.

Do you even read, what people write? There is no need to buy anything from Intel directly. You can buy those SoCs from resellers (even in single quantities). Intel would have to stop selling their chips to all resellers, which is extremely unlikely to happen. Why would they do such a thing?

When choosing the right SoC, there are a lot of things to consider. That's for sure, but this is not one of them.

 

[WizardStan]

Kind of is. If he wants as much buys as possible.

He doesn't, he's already said that with good reasons. Big companies with large support teams want many sales; small companies with very limited support like DragonBox need to straddle a very careful line between maximizing profit and selling more than they can handle in terms of supporting active units and handling returns.He has a very legitimate fear that if he tries to go x86 demand will exceed what his company can currently support, both in terms of gross numbers and an increased percentage of returns for reasons of "it doesn't do what I thought it would". I don't think the number will be nearly as high as he seems to be afraid it might be but it will still not be zero, I've seen people complain about multi-thousand dollar Apple laptops that couldn't handle desktop like graphics in Windows games, people that just didn't do their research before making a purchase, and I have no reason to believe that those same people won't jump on a $600 palmtop that advertises the ability to run some Windows games natively.

At first I was being swayed by arguments about the capabilities of x86 (64bit multi-arch trauma aside) but the more I think about it the more I'm siding with EvilDragon on this one, x86 does carry a lot more risk than ARM and I can appreciate why he would hesitate to take that risk right now. Grench makes compelling arguments but they mostly come down to "it won't be *that* bad", which is fair to say when it isn't your money on the line, and I can totally respect ED being scared of the potential, regardless of how "not *that* bad" it might be.

Maybe if the Pyra is successful and he's able to grow his company slowly and controlled he'll be big enough to make an x86 handheld in 4 or 5 years, but right now I think he should avoid unknown risks, especially when he's still carrying debt from the Pandora. Let him recover from that before he starts taking bigger risks with money he doesn't have.

 

[Saber]

Merry-go-round.

Reminds me of a song I like.

Due to the time crunch they'll likely be using the Omap5432. More than capable enough now and later for the majority of the community and new members. Impatiently waiting for the Kickstarter/preordering process.

Hope to see an atypical color for the Pyra. Midnight blue or crimson perchance?

 

[xiongxioi]

Intel would not stop their production line, but they can refuse to sell us any socs some time down the line. Pyra is too small a project for them to care.

Do you even read, what people write? There is no need to buy anything from Intel directly. You can buy those SoCs from resellers (even in single quantities). Intel would have to stop selling their chips to all resellers, which is extremely unlikely to happen. Why would they do such a thing?

When choosing the right SoC, there are a lot of things to consider. That's for sure, but this is not one of them.

If that's true then why can't Ed get anything from samsung? It's my fear that Intel could someday just say screw those pandora hipsters, and we would be getting bugger all.

There's a potential market for x86 umpc I think, but whether we can get a slice is up to debate. Making a product is one thing, marketing it to appeal more people is another. Ed barely has enough money to prototype the pyra with the current setup, I don't see how we can make pyra explode without some miracle occurring.


It should also be mentioned that, project alternation is usually seen as the last measure. If we keep changing this and that, investors may lose their confident. Suddenly changing pyra to a x86 umpc would be like turning a cruiser around 360, even we are still at the early stage, this would not look good on reports. Making sure the project's staying on track and within budget is all ed needs to care about. Switching to x86 is changing the scope of the entire project, it's rather serious as there's less information, guidelines, and experience can justify it. Unless Ed's starting a new line of x86 product, it would be rather awkward for him to explain. This is why most companies would rather push out flawed products than delay the launch and spend time refining them. Cough apple map cough. Sure Ed's investors are kind enough to give him a lot of freedom, they still need their money and profit back. Remember icp 2? Adding backlit sounds simple but it can chain react into a total disaster.

 

[Klumpen]

Having a reseller in between is too expensive.

 

[wej]

If that's true then why can't Ed get anything from samsung? It's my fear that Intel could someday just say screw those pandora hipsters, and we would be getting bugger all.

If? There is no question about that. Just go to mouser and search for the chip. You can buy it right away. What has Intel to do with Samsung anyway? When Samsung does not sell its SoC through resellers, you obviously cannot buy it there.

Having a reseller in between is too expensive.

Such a statement is wrong per definition. Some things might be too expensive through a reseller, other might not. It is perfectly normal to buy electronic parts through resellers. You don't want to source every single part from the actual manufacturer, if that's even possible. That would be a pain to do. Especially when buying in small quantities, it might actually be even more expensive to buy from the manufacturer directly (again, if that's even possible).

 

[Neelix]

As I understand it, the OMAP comes with support from TI.   Does Intel offer similar support to companies who but their chips through a reseller?  does the reseller?  how does that work?

- Neelix

 

[Hồng Thất Công]

It is not a technical issue that the PYRA can't be x86.  It's all about risk that PYRA can't be x86 this time.  You all want x86, but are you all willing to support it?  I am pretty sure notaz will probably quit helping ED with low level kernel stuff.  Which one of you is willing to step up and take notaz's role?  Many current devs will quit as well,  which one of you will donate your free time or have the knowledge to optimze software for the PYRA?  Yes, it sounds as if PYRA depends solely on a handful of people for technical support, well because it is, just like the Pandora has been.  Without knowing who will stay on board to support the PYRA and x86 is an unknown world, it all boils down to risk. 

 

[wej]

Of course Intel offers support for their products. It does not matter where you buy the chips eventually. They even offer things like analyzing your schematics.

Many current devs will quit as well

Maybe so, maybe not. You probably all remember that ED has recently been presenting the Pandora (and the Pyra) at FOSDEM, which is a pretty large open source event with lots of skilled people around, with many of them not (yet) being part of this community. With the Pyra being an interesting device for Open Source developers, it should certainly be possible to get some of them interested in developing for the Pyra (including low-level stuff like kernel development), especially if its hardware is open source friendly.