The 'Raspberry Pi' thread

https://www.raspberrypi.org/documentation/configuration/config-txt/memory.md
I wouldn't be trying to do 4K contetin 64MB of video memory, set it to 1024MB.

 

GumbyNoTalent you directing that to me ?
I only have 1024 total, so setting all of it for the gpu probably isn't going to end well
I have zero idea what kodi/librelec sets the gpu to. I'd assume its 'just enough' as i didn't see a droppped frame during play back. silky smooth playback in kodi was highly impressive

On another note it does seem the limited ram is the cause of the network speed tests being all wonky.
Speedof.me's new test just breaks and the old page is very limited for downloads. On smaller sample sizes I did see 85mbit before the pi would succumb to the amount of data. I suspect if i was writing this off to SSD rather than uSD i might get better results.
Using ookla speedtest i saw 85 down again briefly before dropping to a net ~45mbit and a really good sustained 38 up. A more lightweight OS setup with a SSD to drop the files onto will get me the numbers I'm after (100/40)

and even if it doesn't workout as a torrent box, worst case scenario for this 1 gb Pi is a awesome retropie upgrade for my bartop arcade machine

 

Install htop instead. It gives clearer output.

The orange section of the RAM output is cache, which can be considered available to the system, and not a part of the consumed memory. It can be discarded almost instantly.

Linux aggressively caches file systems to improve performance. No point having RAM sitting around empty, after all. May as well use it to speed things up while applications don't need it.

 

revision hinted at due to USB-C shenanigans.

https://arstechnica.com/gadgets/201...ct-usb-c-design-wont-work-with-some-chargers/

 

Yes let's reinvent the standard because we can, jeez that's irritating

 

Is it the end of the world that a handfull of PSUs are not compatible ?
Really, what do people expect from a US$35 sbc while MSI are selling a USD800 Ryzen x570 motherboard ?

No doubt a class action law suit will be started in the USA.
Oh the butthurt is unbearable....

 

So my Toyota shouldn't have to work with round tyres because Lamborghinis exist? What?

 

not sure on the analogy ? wouldn't it be more a case of my Ssangyong was going to take any unleaded, but now it will only use BP, Shell, Puma, Mobil, Caltex, Liberty, but not United or Pacific

 

Not following a standard and leaving a device not working with a decent selection of standards compliant chargers is annoying. I can't see a good reason why they wouldn't, unless saving a single resistor is that important.

 

Obligatory XKCD:
https://m.xkcd.com/927/

(Read alt-text for surprisingly relevant lols)

 

You may also be shocked to know that most IEEE ISO Standards are written in such a way that many interpretations can be compliant, I have come to this conclusion having spent the past 18 months dealing with RFID.

 

I forgot to check the alt-text like a sucker first time around, ouch that's relevant lol

 

Not the "end of the world", no. But USB is a standard (the hint is what the "U" stands for), and standards should be followed.

To be fair, it's more than "a handful" when mobile phone chargers dominate the USB "PSU" space. But again, it's more that something so standardised should be followed precisely to avoid situations like these. That's kind of the point of standards.

It's also why I don't rush out and buy hardware day one.

All good though. They've acknowledged it and are planning a revision to fix it, so I'm happy with that.

 

Mind you it's not exactly new that people suck at sticking to the usb-c standard. Cables in particular are a bloody clusterfuck all round

 

Trying to get engineers and developers to comply with standards is literally my career.

Technology is easy. People are a mess.

 

http://mloduchowski.com/en/blog/raspberry-pi-4-b-pci-express/

Hmmmm

 

Wooo, Glad I waited.... just like I waited for cat6e...... nods.

 

That is proper impressive!

 

Obviously not buying until this is fixed. Makes you wonder how 'extensive' their field testing really is.

 

Despite their popularity, they're not a huge organisation. Resource constraints are a real thing, and their whole design is not to profit to ridiculous levels, but to provide a cheap, versatile computer to everyone, regardless of geography, social situation or income.

The test phase likely provided people with power supplies, or people used existing units. Smartphone chargers with chips in then trend to be the domain of a different demographic.

I'm not trying to make excuses for them. But understand that they're literally a charity, and as such aren't anywhere near as well set up as giant corporations. The fact that they exist AT ALL in this shitty, greedy, late capitalist world is amazing. (Google projects like "One Laptop Per Child" as an example of a similar charity that crashed and burned).