GPD Win (x86 Computer / Palmtop)

[Blue Shadow]

So something like this? (aspect ratio about the same as the pyra)

If the angles would be flatter (like in your car picture), it could easily either get sharp or would waste space, if you add more edges, it might look rounded again

 

[rygD]

That is nice. I wouldn't mind that. You just round the angles slightly so it isn't so sharp and probably fragile.

If the curves on the bottom of the Pandora were used on the top I think it would look better (and more like later DSs). I think I mentioned that idea before, but someone didn't like it for some reason. It may not work so well with the fat Pyra. I am sure what we get will be fine, even if it isn't pretty. Just pretend it was inspired by an upside down original GB (more curved on some of the top edges instead of the bottom).

 

[fusion_power]

So something like this? (aspect ratio about the same as the pyra)

View attachment 28254

If the angles would be flatter (like in your car picture), it could easily either get sharp or would waste space, if you add more edges, it might look rounded again

Something like that, but where the edges are a little bit rounded in between the transition from vertical/horizontal to the bevel. bevel itself can be flat without curves.
I made a sketch and used the actual Pyra Profile for that. So the size is more or less the same, of ocurse the internals may still not fit anymore due the more restricted edges.

 

[vcoleiro1]

I absolutely love the design of the GPD WIN. In fact that was my first comment about it after seeing it for the first time. The styling reminds me of something from 2001 a Space odyssey , a kind of retro futuristic look. Particularly the first images they showed which was white with a black interior. They still plan to make this version as an option after it releases.

 

[mius]

I would rather have the handheld version of this (or maybe the 68, for the over/under headlights, I haven't decided):

Mah choice:

 

[rygD]

So the size is more or less the same, of ocurse the internals may still not fit anymore due the more restricted edges.

Would you be ok with the bottom being somewhat flat, like the bottom of a car? Then the internal stuff doesn't need to get changed. It still looks good.

 

[levi]

Wasn't a curved top with a flat bottom his original critique of the Pandora?

 

[vcoleiro1]

Only 4 days left in the GPD WIN Indiegogo campaign. Looks like a lot of procrastinators are now jumping in.

 

[rygD]

Wasn't a curved top with a flat bottom his original critique of the Pandora?

No, it was that the curved top doesn't look like it belongs with the less curved bottom. The lid doesn't really fit with the feel of the overall design, which I agree with. If you have angles things tend to work better.

 

[vcoleiro1]

Here's some more photos of the GPD WIN (current state) prototype showing it's size vs the GPD XD. Pretty much the same size

Click on the images to see full size images.
























































Bottom




Top

 

[SNESFAN]

smaller than I thought, looking really good

 

[Chip]

I was thinking it would be slightly larger than the XD, but if anything, it looks slimmer. I've been on the fence this whole time, and now I'm really leaning toward backing this.

 

[vcoleiro1]

@Chip

Well, you better decide soon, only like 3 days 20 hours left

 

[_jr_]

So, what would your solution be?
Sell devices telling they can run between 800MHz and 2GHz, depending on what you do?

Something like this. It is fine, that the system may overheat under certain circumstances, bit it has to be absolutely clear to the customer what to expect. The current information on https://www.pyra-handheld.com/boards/pages/pyratech/ isn't good enough, because it contains not even a footnote about potential issues (like overheating, missing drivers, non mainstream drivers, etc.). It would be nice to have a couple of benchmarks with some performance figures (raw power, battery runtime, thermal considerations) for the customer to understand the capabilities of the device, e.g. some productivity suite (hopefully with a result of 'I can edit graphically heavy Pyra flyers for 20 hours until the battery is empty without the device getting noticably warm), a non optimized FPS port with a demanding level (might give 'Doom 3 runs at 15fps for 6 hours while the device gets a little hot, but doesn't have to be throttled'), a kernel build (could be 'full kernel build takes 1h and drains 10% of the battery without the device getting warmer'), and a stress test type program (perhaps 'with the stress test running the device becomes too hot to handle after 5 minutes while draining 25% of the battery in the process')- perhaps each additionally with the modem with good connectivity working on some torrents in parallel.

'The user just has to understand it properly' should be below your standards unless you provide adequate information from which the user can make an informed decision. Currently even owners of a OMAP5 devboard can't give a good estimate of the actual Pyra capabilities (besides raw power).

 

[EvilDragon]

Of course such infos would be nice, but they are unknown until all that has been tested, right?

 

[levi]

It is fine, that the system may overheat under certain circumstances, bit it has to be absolutely clear to the customer what to expect.

To be accurate, are you okay with the system going over it maximum heat rating and causing electrical errors at least, or do you instead mean that it may reduce clock/voltage etc. as it approaches some maximum rating, avoiding any temporary or permanent damage. I'd not be happy with the former, but more okay with the latter provided I know about it in advance.

 

[Chip]

I think the issue that _jr_ has isn't what the device does when it gets too hot, but that the device gets too hot in the first place.

There is an expectation among consumers that a computing device rated up to 2GHz will be able to actually run at 2GHz indefinitely. This has certainly been the case for desktop and laptop systems. Those devices have the luxury of space for active cooling solutions, so they have no issue exhausting waste heat efficiently.

But handheld devices are a different story. Cooling options are limited, both by the small size of the device and the fact that since you hold it in your hand, it can't be allowed to get hot on the outside. Up until recently, mobile chips simply weren't powerful enough to generate much heat. But consumer demands for more powerful phones and tablets have created a situation where we have chips generating 2W+ of heat and no way to get rid of it.

The answer is colloquially known as "thermal throttling". It's actually a combination of behaviors to keep the chip at safe temperatures. It starts with only running the chip as fast as is absolutely necessary to process the current processes in a timely manner, and ends with forcibly slowing down if it gets too hot, regardless of the processing demands. This isn't new. Smartphone chips have been doing this for years, but because most people don't overly-stress their phones most of the time, it goes unnoticed.

What consumers need to understand (and manufacturers need to explain) is that a mobile device that slows down due to thermal throttling isn't defective or poorly engineered. It's actually working as intended. Physics is physics, and there is only so much you can do to cool a small device without a fan. Manufacturers aren't being deceptive by quoting boost clock speeds that can only be maintained for short periods - that's why it's called boost. If it was the max stable speed, they'd just call it the speed.

 

[Klumpen]

I'd have no problem with a device that's 1-2cm thicker for proper cooling although my use cases don't need high CPU clocks anyway.

 

[levi]

Intel have made thermal throttling a sales benefit, with their 'Turbo mode' rating on chips for a few years now. I don't know if those chips are able to run at their non-turbo clock speed with a working cooling rig, but that's the impression the marketing gives. I don't know if the CPU that GPD are using has those two ratings, and if it'll be able to sustain the lower of the two with the cooling solution applied, but it's perhaps a way to be able to market the maximum rating with a stated indefinite clock rate given also.

 

[fusion_power]

Would you be ok with the bottom being somewhat flat, like the bottom of a car? Then the internal stuff doesn't need to get changed. It still looks good.

I've taken the opportunity to alter both sides. Grip may be just better if the base unit is somehow smaller at the sides. And it looks more symmetrical imho. From ED's transparent Case pics I've spotted some unused space on the bottom (sides) of the Pyra. This could be an advantage for such an design...in theory of course. ^^"

No, it was that the curved top doesn't look like it belongs with the less curved bottom. The lid doesn't really fit with the feel of the overall design, which I agree with. If you have angles things tend to work better.

Both, actualy. Bottom is to straight and lid to curved imho. The Pandora/Pyra is compared thick so it makes also sense to have a more curved bottom line. Devices like the NDS are overall thinner, they can live fine with less curved edges and still look good. Pyra would not work this way for sure if we have a non sloped/curved top and bottom for the sake of symmetry.

The GPD WIN looks so elegant because it has a compared low profile and additional onto that a curved bottom/top. This make the device look even thinner than it actualy is. I wish the Pyra was at least a little bit more thin and/or sloped at the edges for the "illusion" of a thinner device. It may be even OK if the bottom would be as strangely curved as the lid, at least better than just one side.

Intel have made thermal throttling a sales benefit, with their 'Turbo mode' rating on chips for a few years now. I don't know if those chips are able to run at their non-turbo clock speed with a working cooling rig, but that's the impression the marketing gives. I don't know if the CPU that GPD are using has those two ratings, and if it'll be able to sustain the lower of the two with the cooling solution applied, but it's perhaps a way to be able to market the maximum rating with a stated indefinite clock rate given also.

There is something new from the mobile passive cooling-front. Acer brings new ultra portable laptops with fanless "liquid cooling" solution:
http://www.notebookcheck.net/Acer-unveils-Switch-Alpha-12-2-in-1-with-liquid-cooling.164109.0.html

They cool down the Skylake-U Processors (15W TDP) with that thing, it should be more than enough for some 2-4W SoC's then.