That took a while...


@hns that means the pyra can run stable in 2GB mode, even when there is a 4gb chip soldered on it?

So far we have no indication of the opposite.

Is there any stability or speed difference between 2GB as four 512MB Vs 2GB as four 1024MB running @ half capacity? May be too early to know?
[doublepost=1496520798,1496520210][/doublepost]
Did someone just mention 8GB? :p
[joining your sarcasm]
Meh, peanuts! I have 48GB on my Linux desktop machine and my browser really flies! But I'm considering adding a 2nd CPU so I can activate the 2ND bank and go to 96GB to make more betterer! [/sarcasm]

Is 4GB a bad thing? Not really.
Is 4GB a must? Not really.
Is 2GB adequate to the SoC? Yes.
 
The 4GB memory chips work fine in standard 2GB software mode on all prototype CPU boards I have tested so far.
Only trying to make 4GB available in U-Boot (different EMIF config) and use it in kernel (LPAE enabled) raises issues in kernel.
The kernel boots, shows 4GB being available but runs instable.
Which might either be
- pure software (imprecise memory controller config, wrong virtual memory management, wrong cache management) or
- hardware (more noise, more power demand, more whatever, as soon as the second bank inside the 4GB chip becomes active).
Further tests to find out are planned but not yet done.
So turning 4GB off and using 2GB is just using the standard 2GB U-Boot and Kernel. No need for a hw switch.

Sounds serious. Finding RAM-releated issues can be a nightmare, I know that from the PC. ^^""" Let's hope it's not that complicated in case of the Pyra.
 
thanks for this nice video-update!
i'm ok with any color, as long as it is coated with c64c light-beige. :)
 
Is the Pyra still 3.2cm thick? It looks thicker in the Video.
haha, don't be silly... it is a πra... so it should be 3.14159 cm thick...

In fact, if you divide the width by the length, you get the golden ratio...

LET PHI => 1.6180339887498948482
13.9/PHI = 8.59067244362354 (real length 8.7cm, 0.1 cm off from the golden ratio).
 
Last edited:
I so want to rip the foil off those displays…
Agreed. When I first saw those images I thought he might have cracked displays during assembly. It took closer inspection to realize it's just the peel-away shipping & assembly scratch guard layer.
 
Back
Top