Chris Ellis @intrbiz@bergamot.social

@xahteiwi @JetlagJen

I do think it's unfair to say 'conveniently excluded'. That would imply a lack of effort, which is not the case.

Both local and national govt, do put effort into reminding people, reaching people and making it easy.

Often political parties, schools and other activists groups also remind people.

Short of compulsory national id cards, I don't see there is much more which could be done to get people on the electoral roll.

@regordane @xahteiwi @larsmb

The UK has a very laxidazical approach to identity.

In many ways each govt department has their own approach to identity and often their own reference numbers.

The concept of a national id scheme was floated under Blair, but was very unpopular, for many reasons.

As such the election system effectively maintains it's own list of people, as does many other govt departments.

@xahteiwi

No public stats that I know of, some might be possible via FOI requests.

In people are reminded yearly to keep it upto date. It's kind of linked a bit with council tax, so in general it's more the 'head of the household' having to notify of other residents whom are eligible.

It's more akin to having to fill in a census in my view.

I suspect it's often students and multiple occupancy rentals that mostly fall through the cracks.

In many ways our lack of govt. id makes this harder.

@xahteiwi

People are not required to vote, but are required to register.

In practice the council send a letter to each property asking people to confirm their details to be added to the register. In the past some councils would have people going door to door to ensure the register was upto date.

It's also worth noting the local council collect and count votes, for both local and national polls.

I've never heard of someone being fined.

@jawnsy

The stock image was running Kernel 6.1, which predates your LKML post, so would depend if it was backported.

I want to have a play with UEFI support and get mainline booting when I get a chance.

@jawnsy @yawnbox

Not sure if any acceleration was in use. 50% perf hit for encryption is a bit more than I expected TBH.

In general the IO throughput seemed better than I expected. Easily beats an RPI CM4 with an SSD.

Was testing pretty quickly, might play a bit more if I get some time in the week.

@jawnsy @yawnbox

IO Benchmarks:

IO, 1 job:
write: IOPS=239k, BW=933MiB/s
readwrite: read: IOPS=128k, BW=499MiB/s; write: IOPS=128k, BW=499MiB/s

IO, 4 jobs:
write: IOPS=686k, BW=2681MiB/s
readwrite: read: IOPS=453k, BW=1770MiB/s; write: IOPS=453k, BW=1770MiB/s

DM-Crypt IO, 1 job:
readwrite: read: IOPS=75.7k, BW=296MiB/s; write: IOPS=75.7k, BW=296MiB/s

DM-Crypt IO, 4 jobs:
readwrite: read: IOPS=235k, BW=918MiB/s; write: IOPS=235k, BW=917MiB/s

@jawnsy @yawnbox

I did some benchmarking of the Rock 5B, stock install of the provided OS. Power stats taken directly from lab bench PSU. Power looks to easily be under 20W for most applications.

Rock 5B (16GB) with Crucial P3 512GB SSD, running Linux rock-5b 6.1.43-14-rk2312.

Power:
Idle: 5V @ 0.66A
All CPU stress: 5V @ 1.39A
Single CPU Stress: 5V @ 0.81A
7Zip Bench: 5V @ 2.18A
CPU Stress + IO RW: 5V @ 1.69A
IO 1 Job RW: 5V @ 1.32A
IO 1 Job W: 5V @ 1.71A
Multi-job IO: 5V @ 1.71A - 2A

@yawnbox I suspect for your usecases it would be <= 20W.

It's probably being cautious on what peripherals people attach.

AFAIK it's the same CPU as the Orange PI Max.

I can measure the power consumption of mine over the weekend, if you want?

@dnavinci @ChrisMayLA6

Yes, I would agree pay is also behind where is needs to be. Especially the gulf between UK and US pay for the industry.