SoC Poll - What's most important for you?

Power or Battery - what's more important for you?

  • Power - I don't care about battery, as long as it's the FASTEST!

    Votes: 38 11.0%
  • Balanced - I don't mind a little shorter battery life if it has more power

    Votes: 199 57.5%
  • Battery - I'm pretty sure every current SoC is fast enough for my needs, but I want a long battery l

    Votes: 109 31.5%

  • Total voters
    346

Also note that if you want it to legally run Windows, you probably have to sign some kind of exclusivity contract with Microsoft that disallows you from shipping or supporting any other OS. Of course we can still all install and unofficially support GNU/Linux, but not without paying for Windows first. At least that's how I understand things.
 
It's pure FUD you know...

I can guess you never owned non-branded desktop computer or even built one yourself otherwise i'm assuming you're spreading this FUD on purpose.
I owned plenty of branded and non-branded x86 PCs, starting from an 8088. I don't think you can compare mostly standardized desktop PC hardware to a niche handheld with lots of custom parts. Do you really think that people will be able to just buy a license for a regular desktop version of Windows and it will work out of the box on the Pyra?
_wb_ IF ED wanted to natively install Windows and ship the device with Windows and get a discount on those licenses, then yes, they may have an exclusivity requirement in that agreement.  You're trying to use fear again to twist the argument.

None of us want to use Windows as the primary OS on the hypothetical z3770 x86 Pyra running 64bit Debian.  If a device owner chose, of their own inclination and cost, to purchase a retail Windows install DVD; Windows would, in theory, load directly into a Virtual Machine on said machine, straight out of the box.  IF a sub-segment of the device owner population chooses to install Windows as the base OS on THEIR OWN HARDWARE- and works out how, they're welcome to self-support on Windows.  That's part of the joy of having an 'open' device.

The UEFI bootloader everyone seems to fear is a non-issue.  Which UEFI to use gets chosen by the device manufacturer.  See the Intel example using a 64bit UEFI on this machine:

http://www.phoronix.com/scan.php?page=article&item=intel_baytrail_nuc&num=1

That Intel NUC uses another SoC in the BayTrail family.  From a UEFI, video and processor level it's the same as the z3770.

Wow there are a lot of half truths and fear mongering going around on this.
 
Can we please bury this whole x86 and Windows stuff and focus on which ARM SoC to use?

These threads are getting completely out of control!

I think that's one reason why ED decided quite early to go with OMAP,

although Snapdragon still seems to be an alternative, but I don't know enough about the influence on coding and battery life to say that for sure.

QEMU + WINE would be the way to go for me for running Win-Classics if I would want that at all, but besides from that, I don't want to have any trace of Microsoft in my device, period.

(Yes , QEMU + Windows Installation is also out of question for me).
 
Last edited by a moderator:
Unfortunately interfaces within kernel tend to change over time
Yes, that is a thing that might one day happen, and if it does we change the module to accommodate. We'd have to do this anyway regardless of whether the module were in mainline or not.
'We' means 'there should be someone,right?' or you know someone who will do it?

Or maybe you'll do it yourself? Fine by me then
 
Well again, 

Also note that if you want it to legally run Windows, you probably have to sign some kind of exclusivity contract with Microsoft that disallows you from shipping or supporting any other OS. Of course we can still all install and unofficially support GNU/Linux, but not without paying for Windows first. At least that's how I understand things.
 
It's pure FUD you know...

I can guess you never owned non-branded desktop computer or even built one yourself otherwise i'm assuming you're spreading this FUD on purpose.
I owned plenty of branded and non-branded x86 PCs, starting from an 8088. I don't think you can compare mostly standardized desktop PC hardware to a niche handheld with lots of custom parts. Do you really think that people will be able to just buy a license for a regular desktop version of Windows and it will work out of the box on the Pyra?
_wb_ IF ED wanted to natively install Windows and ship the device with Windows and get a discount on those licenses, then yes, they may have an exclusivity requirement in that agreement.  You're trying to use fear again to twist the argument.

None of us want to use Windows as the primary OS on the hypothetical z3770 x86 Pyra running 64bit Debian.  If a device owner chose, of their own inclination and cost, to purchase a retail Windows install DVD; Windows would, in theory, load directly into a Virtual Machine on said machine, straight out of the box.  IF a sub-segment of the device owner population chooses to install Windows as the base OS on THEIR OWN HARDWARE- and works out how, they're welcome to self-support on Windows.  That's part of the joy of having an 'open' device.

The UEFI bootloader everyone seems to fear is a non-issue.  Which UEFI to use gets chosen by the device manufacturer.  See the Intel example using a 64bit UEFI on this machine:

http://www.phoronix.com/scan.php?page=article&item=intel_baytrail_nuc&num=1

That Intel NUC uses another SoC in the BayTrail family.  From a UEFI, video and processor level it's the same as the z3770.

Wow there are a lot of half truths and fear mongering going around on this.
Grench, like Ed stated, it's all about unknown risks his company has to face going x86.  It's not a technical issue.
 
Last edited by a moderator:
There are really good at the moment oses replaements for windows on x86: haiku os and react os alternatives. Even freedos. Haiku is already best choice. Yes its sometimes unstable but its awesome choice for windows haters
 
'We' means 'there should be someone,right?' or you know someone who will do it?
Yes, "we" means "someone, somewhere", specifically someone associated with the Pyra project. This is true regardless of whether the entire thing end to end is open source or if only the kernel interface is open source. The fact that part of the driver is a binary blob does not change the fact that if the kernel interface changes for any reason that someone somewhere will need to do work in order to bring it up to date.
 
It could've very well been that the same people would be in favour of an x86 SoC so they can also play normal x86 games (both Indie, commercial ones and Steam), etc.
What's wrong with that? Expanding the capabilities shouldn't be a drawback.
However, there are dangers as well:
Fear the dangers...
* Maybe no one would release optimized games anymore (screen size, button config, etc.) because everything exists for x86 anyways, so why re-port it?
With a proper Linux driver for the nubs (joystick 1 & 2), game controls (game pad 1) and shoulder buttons (joystick 1 button 1 & 2), they wouldn't NEED to be ported. Screen size fears are laughable. Just play at a lower than native resolution and pop the fonts are readable. This is another case of trying to twist a benefit (a 5" 1920x1080 screen) into a fault (OMG it'll be unreadable!) and then blaming it on something entirely non-related (x86 Games).
* People thinking you can run all Windows games and then getting angry because they don't work
That is the elitist, "x86 would bring in dumb smelly peoople," argument again. In any large population, idiots will exist. YOU get to choose whether or not you even want to advertise that it -can- run Windows applications in WINE. IMHO advertise that it runs native X86 Linux applications similarly to a Core 2 Duo machine from 2007 with business class graphics. Ignore the WINE & Windows capabilities in advertising and allow them to be discussed in a 'unsupported subforum'.
Again, none of us who want the z3770 to be considered have any intention of using Windows as the base OS. That is a non-issue.

* Security issues - popular systems (x86) are more probably being hacked or attacked with viruses (I know x86 Linux rootkits / viruses exist, but I don't think they exist for ARM as well (except for Android)
More FEAR. Now you're telling us to be afraid of running X86 Debian Linux? One of the most secure OS platforms on the planet? Really?
Being niche is not always a bad thing - it pretty much ensures you're different from everyone else, which means less attacks onto your system.
More fear. Do you really think X86 Debian is an insecure platform?
 
Can we please bury this whole x86 and Windows stuff and focus on which ARM SoC to use?

These threads are getting completely out of control!

I think that's one reason why ED decided quite early to go with OMAP,

although Snapdragon still seems to be an alternative, but I don't know enough about the influence on coding and battery life to say that for sure.

QEMU + WINE would be the way to go for me for running Win-Classics if I would want that at all, but besides from that, I don't want to have any trace of Microsoft in my device, period.

(Yes , QEMU + Windows Installation is also out of question for me).
Agreed. This thread has served it's purpose and, just like the other voluminous SOC thread, we've reached an impasse. One vote for it to take the long dirt nap please. B)

What's important now is which one?

1. 805

2. Allwinner

3. Omap5

4. MT6595
 
Last edited by a moderator:
There are really good at the moment oses replaements for windows on x86: haiku os and react os alternatives. Even freedos. Haiku is already best choice. Yes its sometimes unstable but its awesome choice for windows haters
How is Haiku is Windows replacement? Just by being an alternate OS? I could see the argument for ReactOS since it's actually intended to run Windows software but not Haiku.. Or are you just saying that there are OSes people could run that aren't Windows?

I suppose it's not as absurd as actually saying people could go from Windows to FreeDOS..
 
Can we please bury this whole x86 and Windows stuff and focus on which ARM SoC to use?

These threads are getting completely out of control!

I think that's one reason why ED decided quite early to go with OMAP,

although Snapdragon still seems to be an alternative, but I don't know enough about the influence on coding and battery life to say that for sure.

QEMU + WINE would be the way to go for me for running Win-Classics if I would want that at all, but besides from that, I don't want to have any trace of Microsoft in my device, period.

(Yes , QEMU + Windows Installation is also out of question for me).
You're clearly confused. X86 (z3770) does NOT mean Microsoft Windows any more than ARM means Android.

Opening up NATIVE LINUX X86 commercial software is what we're after. Steam for X86 Linux. Neverwinter Nights for X86 Linux. VirtualBox for X86 Linux. The entire X86 Debian repository. THOSE are what we're after. There is no practical alternative for ARM that opens up those possibilities.
 
You're clearly confused. X86 (z3770) does NOT mean Microsoft Windows any more than ARM means Android.
No, I am not. Both discussion strangs are annoying in the same way and are connected by people that (in favour of x86) say, being able to run Windows would be a good thing

and if you want to read everything I write like I was dumber than I am, then you are free to do so, but please don't state it as a fact.

If ED would get a cheap devboard for this intel chip and if it would be better than ARM regarding power consuption, maybe I would want him to go there, but both is not the case atm if you ignore what Intel says about their own chips and how much better than anything else they are - they even got no devboards available for him as it seems.

Most of this thread is speculation and gnostic stuff, no data.
 
Last edited by a moderator:
Accusing me of fear mongering might not be the wisest move when you link to something else than actual source. Steam for linux is closed source, so is neverwinter nights. Omap5 has hardware visualization, debian has the xen hypervisor, wine is coming to arm. The entire armhf debian library is just as good a way of saying the entire one, except it makes your point moot. I have been using debian arm library since ARMv5.

And by "we" i suspect you meant something other than what is stated in this thread, a selective discourse for closed source, which you so conveniently labeled "commercial software"

If "native" in your argument isn't redundant it means binary only (closed source) commercial linux software.

In brackets here because commercial does not entail closed.

A variety of which there isn't an abundance to say the least, and the same arguments that favour doing the massive work of porting it to linux, should mean it will land on arm if it isn't there already. For the case of argument in debian, since you are talking about steam specifically.

The UEFI bootloader everyone seems to fear is a non-issue.  Which UEFI to use gets chosen by the device manufacturer.  See the Intel example using a 64bit UEFI on this machine:

http://www.phoronix....trail_nuc&num=1

That Intel NUC uses another SoC in the BayTrail family.  From a UEFI, video and processor level it's the same as the z3770.

Wow there are a lot of half truths and fear mongering going around on this.
Edit: https://downloadcenter.intel.com/SearchResult.aspx?lang=eng&ProductFamily=Desktop+Boards&ProductLine=Intel%C2%AE+NUC+Boards+and+Kits&ProductProduct=Intel%C2%AE+NUC+Kit+DN2820FYK

is the link to where you can download binary only files to flash intels own offering, on a platform that isnt what we are discussing. How is "intel visual bios" to begin with, not an issue?

Besides, every x86 platform is different with a whole heap of inept things making even the most minute change to even a comparable platform needing a new uefi bios implementation. And if you want it to be anything sane, compare, effort spent implementing coreboot support. Bring me the full truth on how vendors are given full access to the documentation needed under non NDA agreements before you start comparing.

So I'm clicking around on said page, and the EFI implementation doesn't seem to write itself. Lets see, what else is on offer, windows only NIC and audio drivers, intel trusted computing, oh, documentation, nice.  Whats this, a license

LICENSE. This Software is licensed for use only in conjunction with Intel motherboard products. Use of the Software in conjunction with non-Intel motherboard products is not licensed hereunder.

First line of said license...

How hard is it to find an open implementation for omap5? ill spare you the trouble  http://git.ti.com/ti-u-boot/ti-u-boot
 
Last edited by a moderator:
I just find it funny that people who favor x86 has been calling people who favor ARM ARM fanboys and not considering themselves x86 fanboys.   :)
 
Last edited by a moderator:
You're clearly confused. X86 (z3770) does NOT mean Microsoft Windows any more than ARM means Android.

Opening up NATIVE LINUX X86 commercial software is what we're after. Steam for X86 Linux. Neverwinter Nights for X86 Linux. VirtualBox for X86 Linux. The entire X86 Debian repository. THOSE are what we're after. There is no practical alternative for ARM that opens up those possibilities.
 Yes, but it seems only 20% so far are after that, whereas 71% rather want an ARM device here.

So you are speaking for a minority, why are you trying to ignore the ones who want ARM?

It could've very well been that the same people would be in favour of an x86 SoC so they can also play normal x86 games (both Indie, commercial ones and Steam), etc.
What's wrong with that? Expanding the capabilities shouldn't be a drawback.
Don't grab my words out of context. I never said this is wrong.

This was a reply to an accusation that voting here is unfair because it's a board full of people loving ARM.

Which is not true in my opinion - the majority of people here like the Pandora for what it is, not because of the architecture.

Still, it seems most favor ARM for some reasons. It could've also been they favored ARM, but that doesn't seem to be the case.

And whether I think it's wrong or not, whether I favor ARM or x86 doesn't matter in this context, because I was simply speaking of the voting results.

Why do you keep repeating the same things all over and over again?

Trying to convince the 61% who rather want an ARM to switch to x86?

Don't you think they READ already what that would open up? Nearly every third post repeats what you can do with x86, still for some reason people here want ARM.

But this seems something you can't accept.

Why?

There are reasons why ARM is better, there are reasons why x86 is better.

But there's no reason to try to force anyone with repeating things to change his opinion.

However, there are dangers as well:
Fear the dangers...
Does that mean you will provide the needed backup money for me in case something goes wrong?

Great :D

* Maybe no one would release optimized games anymore (screen size, button config, etc.) because everything exists for x86 anyways, so why re-port it?
With a proper Linux driver for the nubs (joystick 1 & 2), game controls (game pad 1) and shoulder buttons (joystick 1 button 1 & 2), they wouldn't NEED to be ported. Screen size fears are laughable. Just play at a lower than native resolution and pop the fonts are readable. This is another case of trying to twist a benefit (a 5" 1920x1080 screen) into a fault (OMG it'll be unreadable!) and then blaming it on something entirely non-related (x86 Games).
You know, I bought the first ASUS netbook available, to have a mobile gaming platform.

I gave up, because I tried so many games and was pissed off to configure them all.

Change key config, switch to fullscreen, change the resolution, find out which ones would even be fast enough to even enjoy them, etc. etc.

This was my most annoying experience with gaming I had.

When I go into the Pandora repository, I can install the games and they are simply working.

Configured to use the correct controls, optimized for the screen size and resolution, fast enough to be enjoyable (if not, it's mentioned in the description of the PNDManager), etc.

I first had a C64, then an Amiga, then played a bit of PC - and was horrified.

On C64 and Amiga, all games were the same. On the PC, it depended on the speed, the sound card, the graphics card, etc.

It annoyed me that much that I switched to playing console games only. That hasn't changed.

I tried a few PC games... Gothic4 was too slow for my PC (yay!), Civilization 5 is okay, but I had to fiddle around with settings a lot so it doesn't stutter anymore, etc.

Nothing has changed for me there.

I don't want that experience to happen with Pandora or Pyra customers.

It should be like Vita or 3DS: Grab a game, play it. No fiddling, no hoping it would even run, no anything.

Quality over Quantity.

That might also be a reason people choose to stay with ARM.

Being niche is not always a bad thing - it pretty much ensures you're different from everyone else, which means less attacks onto your system.
More fear. Do you really think X86 Debian is an insecure platform?
I've had to clean up a few Debian servers from rootkits... if people don't secure and update it regularly, this can happen.

Are you thinking everyone will update the Pyra regulary?

Don't be sure. I sometimes get Pandoras for repairs that don't even run SuperZaxxon, but Zaxxon...
 
I know this off-topic and I meant to address this earlier, but someone said Vita and 3DS

are "disposable". Without farther conversation on this, you are off your rocker.
 
I know this off-topic and I meant to address this earlier, but someone said Vita and 3DS


are "disposable". Without farther conversation on this, you are off your rocker.
But windows is more disposable.  Without farther conversation on this, you are off your rocker  :p
 
Last edited by a moderator:
If ED would get a cheap devboard for this intel chip and if it would be better than ARM regarding power consuption, maybe I would want him to go there, but both is not the case atm if you ignore what Intel says about their own chips and how much better than anything else they are - they even got no devboards available for him as it seems.
That is pretty important.

I tried to find out more about the z3770 and power consumption today.

SDP is 2W, but Intel doesn't mention TDP.

With SDP (Scenario Design Power), we don't know how fast it is. It could be SDP is simple webbrowsing, not gaming.

Most people assume TDP is 4W, but no one really knows.

Then I read some tests about the Dell Venue Pro 8 (a tablet that uses the z3770).

Quotes from Engadget:

During gaming and other intensive tasks, the tablet gets warm enough that you can feel the heat even through the folio case. If you're just surfing the web, though, you can expect the tablet to stay cool.
Okay, so it seems to heat up quite a bit. Doesn't say much about power usage (we don't know how it's cooled), but it's an indicator a lot of the power gets transformed into heat with heavy usage.

Dell specifically claims that the Venue 8 Pro's battery can last up to 9.9 hours (that's nine hours and 54 minutes, for all you lazy mathematicians out there). Perhaps with light usage, you can get the tablet to last that long, but in our video-looping test, we managed a more modest seven hours and 19 minutes. Then again, that was with the brightness fixed at 50 percent.
So, 7 hours and 19 minutes with half brightness playing videos.

Playing videos is using hardware acceleration, so the CPU shouldn't do much.

Slashgear reports the same:

With heavy use days that consisted of lots of screen time and lots of WiFi we were seeing a little more than 7 hours.
And WinSuperSite:

With screen brightness on auto, the Venue 8 Pro delivers about 8 hours of battery life in my streaming HD video tests. Set it to something more akin to what the Surface 2 delivers by default and the battery life in those same tests drops to a more middling 5 hours or so. That is a big difference, and while I'm sure there's a happy middle ground that could net, say, 6.5 or 7 hours of battery, it's also less than I had expected and hoped for.
The Panel is an 8" IPS panel with 1280x800 resolution. According from what I read, it doesn't seem to be the fastest (refresh rate) and brightest one, so my assumption is it roughly uses as much power as the Pandora panel.

I can't tell this for sure, but the Pyra panel uses about the same power as the one in the Pandora, but is 5" and FullHD - so I think the 8" one with lower resolution and refresh rate should roughly be the same.

The battery has a bit more capacity then the one in the Pandora... so those 7:19 hours don't really impress me.

Of course, it could be something else in the tablet, but we have no idea.

It's just an assumption - but so is mostly everything else in this thread here.

We don't know yet the power consumption of the ARM chips, so Intel maybe Intel really IS better, but we don't know.

SDP is a pretty rough value - who defines what the typical scenario is? For tablets probably webbrowsing and playing small games.

We know Intel claim it uses less power than other ARM SoCs.

We also know Intel cheated with benchmarking last year to get better results than others.

And doesn't give out a maximum consumption, only a "typical use" one.

And right now, we don't have anything to test it, as no devboard is available.

That is quite risky, in my opinion...
 
Last edited by a moderator:
1367355312888.jpg


Seems I was wrong earlier.
 
@ED

You'd better read review sites with actual measurements and established testing methodology. Support German tech journalism, read notebookcheck :)

http://www.notebookcheck.net/Review-Lenovo-Miix-2-8-Tablet.107505.0.html

http://www.notebookcheck.net/Review-Dell-Venue-8-Pro-Tablet.107892.0.html

http://www.notebookcheck.net/Review-Toshiba-Encore-WT8-A-102-Tablet.108992.0.html

Lenovo Miix has the lowest average power consumption for wifi browsing (2.31 W for all system including the screen at 158 cd/m^2 and wifi which may be more powerful than typical phone wifi). The interesting thing is that Miix have a battery just 13% more capacious than pandora so SoC power is certainly less than 2W in this scenario.

Some ARM tablets for comparison

http://www.notebookcheck.net/Review-HP-Slate-7-Plus-4200ef-Tablet.107523.0.html - tegra 3, 7'', 2.5 W, video playback makes things even worse

http://www.notebookcheck.net/Review-Amazon-Kindle-Fire-HDX-7-Tablet.104415.0.html - snapdragon 800, 7'', 2.4 W. If it doesn't have some super-inefficient screen that means that snapdragon 800 will be able to kill pyra battery in approximately 3.5 hours as this tablet have only slightly better battery than pandora.

You can find more interesting tests there.

Actually as specified in intel datasheets SDP is meant to basically tell that cooling solution should be able to dissipate SDP W when the SoC is at 70 degrees Centigrade.

Besides it's not like the need to very good cooling solutions is something new in ARM world - see Anand's frozen phone benchmarking experiments.

I don't think that current intel SoCs are better performance-wise than current ARM A15 chips, it wasn't meant to be that way actually. But they aren't much worse to discard them as bad options.

The only reason to go intel is support. There is also a high chance that some intel OSTC developers will be interested in pyra and at least test their patches before committing to kernel tree.

Pandora support was always very problematic for me. I couldn't basically use a pandora for about a year until finally 3.х kernel arrived and fixed sd corruption bug. And now we have userspace so old that debian stable seems like something very new :)

As far as I understand currently kernel development is done by one man 'notaz' kernel development team. Pandora is a good experiment and as such any development done by more that 0 people is a good thing but I want pyra to be something more mature. I think depending on a single man is a bad thing for project stability. He can be hit with some Real Life issues and quit or suspend his fork for unspecified amount of time.

That means we have to either extend our development team with some paid developers or maybe just some support contracts as we can surely care of pyra-specific bits by ourselves (it's questionable whether we can develop this bits but we can maintain it for sure) but currently won't have manpower to maintain drivers for generic components like SoC components, wifi and 3g modem.

Another way is to merge with the other community that have such developers willing to do it for their reasons. That may be x86 or probably A80 as Allwinner have quite healthy community that is being strengthened by makeplaylive projects. There is a chance that some good community will be formed (or being formed) around snapdragon due to jolla phone 
 
Back
Top