Analyzing 4GB RAM


Useless? I think it captures very well how many of us feel about it.

I support the Pyra not because of it specs. They are more than sufficient. I support the Pyra because I want it to exist.

In a time where everybody was making identical smartphones, the Pandora has showcased another way forward.
Now there are more competitors out there than ever before, but the Pyra is still leading the way.
It is not about clockspeed, it is about utility. So far, none come close.
 
The soc Pyra has, already gives plenty of power for device of such form factor, you should be more worried whether software can actually take use of that hardware properly.

Wot 'e said!

The last ten years of consumer computing have seen raw processing power increase dramatically, but real-world performance really hasn't.
I had the misfortune to use a generic corporate windows/office installation the other day, and, by the FSM, was it awful‽

Every trivial little action I performed was slowed to a craw by stupid special effects, obstructive UI design and spectacularly unintuitive behaviour. The defaults were wrong, and the manual controls were hidden.

At their best, GUIs present clear options to the user - but the way things are going (Android is just as bad, btw) I feel like retreating into a corner and huddling behind a command prompt.

I'm sure many people here feel similarly!
 
just curious as to what SoC's these would be??
"New" is probably pushing it. The affordable socs still use several years old tech, and are not _that_ big update from OMAP5
For example Rockchip would have rk3288 w/ mali GPU which seems like decent choice. (Used in GPD XD)
There's also iMX8 series w/ vivante GPU (based on ARMv8-A) but it's been delayed already by year or so.
 
and there it is... lol
whats available and what ED would be willing to use as well as what the vendors will sell in a small batch limits what SoC's are suitable for the Pyra.
 
First things first, find a graphics subsystem that has open source Linux 3D drivers.
Mali actually has open source driver, which makes rk3288 quite appealing. However there has been threads past where this SoC has been discussed even by ED, and there are concerns.
Vivante also has reversed drivers which is in much better stage than SGX reversing, but yeah.

Neverthless, this discussion is more relevant after Pyra has been out and used by people for at least year or two.
 
Another primary concern is that modern SOC's generate tons of heat. I for one prefer the passive cooling on the Pandora/Pyra but more modern chips NEED active cooling (GPD Pocket)
 
and there it is... lol
whats available and what ED would be willing to use as well as what the vendors will sell in a small batch limits what SoC's are suitable for the Pyra.

The new SoC has also to be nearly pin compatible, feature compatible, suitable for mobile device in therms of power consumtion, power dissipation, mechanical dimensions. And the most important, it has to be significant better than the old one. This is the reason why nobody can name a replacement for OMAP5.
 
The new SoC has also to be nearly pin compatible, feature compatible, suitable for mobile device in therms of power consumtion, power dissipation, mechanical dimensions.
The CPU board can be freely redesigned, that actually provides a lot of freedom. The required feature compatibility is a rather moot point, I don't see an issue with that one. Modifying the main board for that purpose (e.g. adding some bridge chips/controllers) would actually be rather cheap and not too far off, either way - that's the advantage of using a simple dual-layer PCB. As long as everything else can stay the same...

This is the reason why nobody can name a replacement for OMAP5.
The OMAP5 was chosen because it was pretty much the only choice - mainly for other reasons. Our OMAP5 wasn't even designed specifically for mobile usage - that variant was never released.

It is obtainable in small quantities, TI doesn't force you to only use certified companies and you actually get proper non-Android Linux drivers - those criteria set pretty much every other comparable or more powerful ARM SoC on the blacklist, especially the first one, and so far there hasn't been any compelling x86 chip, either.

I'm actually looking forward to the ultra low power chips that may follow Raven Ridge at some point.
 
I'm curious how the latest RISC-V will do with Linux. If the performances are there, the foss communities might build some nice stuff with it :)
 
I'm curious how the latest RISC-V will do with Linux. If the performances are there, the foss communities might build some nice stuff with it :)

Rough comparison:
U54-MC RISC = 4 * 1.7 DMIPS/MHz
vs.
Cortex-A15 = 2 * 3.5 DMIPS/MHz
 
The new SoC has also to be nearly pin compatible, feature compatible, suitable for mobile device in therms of power consumtion, power dissipation, mechanical dimensions. And the most important, it has to be significant better than the old one. This is the reason why nobody can name a replacement for OMAP5.

this is not true. there are two connectors to the mainboard that has the IO and the Keyboard matrix. there pinout and the locations of them are the only restrictions on designing a new CPU board. the board could possibly be slightly bigger in area depending on what room is available in the case. i think the 3 main parts to this CPU board are the SoC itself, the RAM and the eMMC. plus any supporting circutry required.

, and so far there hasn't been any compelling x86 chip, either.

I'm actually looking forward to the ultra low power chips that may follow Raven Ridge at some point.

problem with x86 based chips, especially if you want to run any form of windows on it , is that it requires a BIOS to be written for it. i believe ED has stated this in one of the CPU threads but i couldnt find the link.
 
Once they got the EXACT cause narrowed down and a solution in mind, all that should be needed is either retrofitting of the old boards, or (more likely based on what Nikolaus has been saying) a new board with redesigned traces. Proto that, test that, if it's all good, make it.

Cases can be made whenever, presumably once ED moves the molds.

I'd say 3-6 months, and that 6 is rather generous. I think it'll be more like 3-4 realistically.

So, your realistic 4 month are over. Where are we? Nearly almost there where we were 4 month ago and far far away from shipping.
 
So, your realistic 4 month are over. Where are we? Nearly almost there where we were 4 month ago and far far away from shipping.

Oh, we are?

Let me see... 4 months ago, we narrowed down the 4GB RAM issue but hadn't done any layout change nor Simulation to make sure the issue has been fixed.
The remaining issues of the main board and case also weren't fixed back then.

Now that has all happened and the CPU boards are already in production since a few weeks, the noise fixed main boards have arrived yesterday and the cases have received all fixes as well (though I need to receive them to confirm that).

Right, absolutely nothing has happened.
Oh, well, but then again, nothing has happened and the the last few years - because a few years ago, we also didn't have working boards or cases. Oh, we also didn't have them 10 years ago... Does that mean absolutely nothing has happened for the last 10 years?

Where has all my time gone?

Sent from my XT1650-03 using Tapatalk
 
Oh, we are?

Let me see... 4 months ago, we narrowed down the 4GB RAM issue but hadn't done any layout change nor Simulation to make sure the issue has been fixed.
The remaining issues of the main board and case also weren't fixed back then.

Now that has all happened and the CPU boards are already in production since a few weeks, the noise fixed main boards have arrived yesterday and the cases have received all fixes as well (though I need to receive them to confirm that).

Right, absolutely nothing has happened.
Oh, well, but then again, nothing has happened and the the last few years - because a few years ago, we also didn't have working boards or cases. Oh, we also didn't have them 10 years ago... Does that mean absolutely nothing has happened for the last 10 years?

Where has all my time gone?

Sent from my XT1650-03 using Tapatalk

By mid March I'm back and have not much at University.
Let's make some Videos then and show them what we got :)

If I had to choose between the Raspberry Pi 3 and my bad Wifi, bad housing non LTE Prototype with the need of booting from my left SD instead of the internal storage I would prefer the pyra.
It feels so much better already.

So, your realistic 4 month are over. Where are we? Nearly almost there where we were 4 month ago and far far away from shipping.

The ptototypes are running for those 4 months.
It is that annoying bug fixing and waiting for boards and parts to arrive that are ordered.
We are on the way.
It is already usable. Only not good enough to sell it.
 
Back
Top