GPD Win (x86 Computer / Palmtop)


First power up image of the GPD WIN board.

7BIBlTD.jpg
 
Last edited:
Afaik Windows can do the rotation, the question is if it is going to have an impact on the performance.
 
Rotation should be handled in hardware by way of the graphics driver. There won't be any performance impact. I mean this is a tablet SoC, and tablets are specifically intended to be used in either orientation. I can't imagine intel wouldn't work that into their hardware and drivers.
 
Well, it depends. I don't know how it works with Windows, but as tablets do it, it will probably work fine.
Would be a problem with games directly using the framebuffer, but I don't even know if Windows even lets you use the framebuffer directly.

The rotator chip would not be needed on the Pyra if you use Wayland (which uses OpenGLES for rotation), but lowlevel framebuffer stuff is an issue.

It could be that Linux would have issues/slowdowns on the GPD.
 
I mean this is a tablet SoC, and tablets are specifically intended to be used in either orientation.

Well, that's not really an argument. The OMAP5 is also a tablet SoC (we even got a heating dissipation guide from TI showing how it will work in a tablet).

Tablets use the 3D chip for rotation though (that would work on the Pyra as well), whereas we also want lowlevel stuff (accessing the framebuffer directly) without the 3D driver chip working.
 
Okay, browsed the internet a bit.
Apparently, some games (especially older ones) ignore Windows rotation settings. Some can be fixed with manually changing then settings in the registry (Warcraft 3), with nvidia, the rotation needs to be set in the nvidia control panel and not in Windows, otherwise many Steam games seem to ignore that as well. Don't know how that's handled with Intel graphics though.

So yes, it could be that some games will have issues, but it's hard to confirm unless trying it yourself, you don't find many posts in the internet about that.
 
Okay, browsed the internet a bit.
Apparently, some games (especially older ones) ignore Windows rotation settings. Some can be fixed with manually changing then settings in the registry (Warcraft 3), with nvidia, the rotation needs to be set in the nvidia control panel and not in Windows, otherwise many Steam games seem to ignore that as well. Don't know how that's handled with Intel graphics though.

So yes, it could be that some games will have issues, but it's hard to confirm unless trying it yourself, you don't find many posts in the internet about that.
GPD probably just thinks "not our problem" like they use to.. ;)
 
Last edited:
Well, ED did the footwork for them, so they only have to find a fix (still the hardest part). Since ED is (was?) a reseller, his customers will thank him for finding that this ifs a potential problem that may be fixable. If there is no official solution I am sure someone will work on it, possibly within our community.

I wonder how many people will buy both a Pyra and a Win.
 
Well, it depends. I don't know how it works with Windows, but as tablets do it, it will probably work fine.
Would be a problem with games directly using the framebuffer, but I don't even know if Windows even lets you use the framebuffer directly.
My PC Monitor has a pivot function, I can rotate the screen image with the drivers or I can keep the screen like before when rotating the Monitor and use the software function if supported (MAME for example). I didn't spotted performance releated issues actualy, however, I mostly used the rotated screen for vertical shooters through MAME and didn't do FPS tests.

Well, that's not really an argument. The OMAP5 is also a tablet SoC (we even got a heating dissipation guide from TI showing how it will work in a tablet).

Tablets use the 3D chip for rotation though (that would work on the Pyra as well), whereas we also want lowlevel stuff (accessing the framebuffer directly) without the 3D driver chip working.
Does the rotator chip actualy has a performance impact on the Pyra? I'm not sure if this was already tested actualy.
 
Does the rotator chip actualy has a performance impact on the Pyra? I'm not sure if this was already tested actualy.
From what I remember, yes. I think the reason the rotator chip was used because it greatly improved performance. I can't remember all the details. Part of the reason for going with HD instead of FHD was how bad it was trying to rotate the image with just what was on board.

I feel like I may be having a language fail with "performance impact". If I misunderstood I will try again later.
 
Does the rotator chip actualy has a performance impact on the Pyra? I'm not sure if this was already tested actualy.
IIRC, ED mentioned that the rotator chip has just enough internal RAM to store a single frame (well two actually, the original and the rotated one). So it is impossible for it to introduce any delay beyond 1 frame. He also mentioned that the rotator chip is fast enough to do all it's work in between frames, so every single frame should make it to the LCD as well (no frame drop).

The only thing that's I havent heard news about is the driver to make the rotator chip work. Last I heard is that hns was working on it.
 
Part of the reason for going with HD instead of FHD was how bad it was trying to rotate the image with just what was on board.
Well beside the performance benefits, the other reason HD was used instead of Full HD is the Solomon chip only handles rotation up to HD resolutions, but not at Full HD... The chip also has some scaling properties that may be handy as well.
[doublepost=1463083150,1463083027][/doublepost]
The only thing that's I havent heard news about is the driver to make the rotator chip work. Last I heard is that hns was working on it.

Well last I know it's working, but still using a very hacky script.
 
Seriously, even with the rotator chip being able to handle FullHD, I wouldn't have used it after seeing how small everything is. GIMP needed extra manual changes in the configuration files (simply choosing the largest icons still was too small), and many Linux programs simply ignore the DPI settings of the desktop (sadly).
 
Seriously, even with the rotator chip being able to handle FullHD, I wouldn't have used it after seeing how small everything is. GIMP needed extra manual changes in the configuration files (simply choosing the largest icons still was too small), and many Linux programs simply ignore the DPI settings of the desktop (sadly).
Well I hope it will work at the end with the screen scale. I need big fonts through all apps actualy (weak eyes). It was very inconsistant on the Pandora and even the 800x480 screen gave to tiny fonts imho. Of course it depends on what program is used, many (newer) PC games are not made for tiny screens and when ported to an Handheld, it can have unwanted effects.
But at least an modern mobile-OS should handle the fonts scaling fine through all (standard) apps. Scaling on tiny High-Res screens should be actualy no problem since we have "retina displays" and still readable fonts. Otherwise everyone would need magnifiers for his Smartphone. :D
 
Well I hope it will work at the end with the screen scale. I need big fonts through all apps actualy. It was very inconsistant on the Pandora and even the 800x480 screen gave to tiny fonts imho. Of course it depends on what program is used, many (newer) PC games are not made for tiny screens and when ported to an Handheld, it can have unwanted effects.
But at least an modern mobile-OS should handle the fonts scaling fine through all (standard) apps. Scaling on tiny High-Res screens should be actualy no problem since we have "retina displays" and still readable fonts. Otherwise everyone would need magnifiers for his Smartphone. :D
Have you ever used Android on a 27'' screen? It's like a children's book slammed in your face.

In the end it depends on for what it's intended and Debian wasn't intended for tiny screens with high resolutions, so I guess a bit customization is needed just like it is with Android on a big screen.
At least some programs seem to look up the font size in the desktop environment.
 
Back
Top