Homatics Box R 4K Plus

[boxr4k]

I can’t test CE-NG anymore but I can attest similar speeds on CE-NO, at least with the SanDisk on my DuneHD BoxR 4K.

I love hdparm as a tool but definitely not for benchmarks. 3 seconds is way to short to see anything useful and there’s no way to increase the timing to have longer runs. I tend to prefer dd + pv as it allows longer reads so we can see if it remains constant or jumps a bit up and down.

$ dd if=/dev/sda count=6291456 status=none | pv -W > /dev/null
3.00GiB 0:00:27 [ 112MiB/s] [ ... <=> ... ]

and it often reflects real world experience, for contiguous linear reads of big files of course. Random reads are another story

[YadaYada]

@goober

Do you think a smaller capacity stick would be as fast? I know there are differences based on capacity in general, but I’m not sure how extensive those differences may be. Regardless, it’s a cheap enough experiment I suppose.

I doubt it, usually the larger capacity drives of the same line are as fast or faster than the smaller capacity drives. My guess is that USB drives with +200MB/s read rates rely on some feature absent in the 4.9 kernel USB stack.

[mnd]

So I got curious, and got this result:

CoreELEC:~ # dd if=/dev/sda count=6291456 status=none | pv -W > /dev/null
-sh: pv: command not found

Am I doing something wrong, or CE NG doesn’t support this?

[Portisch]

Speed test:

X96 X4 - S905X4 based box - any success?

OK, I did the tests in CoreELEC USB Write Speed (Patriot Supersonic Rage XT - CoreELEC) CoreELEC:~ # dd if=/dev/zero of=test.tmp bs=1M count=1024 1024+0 records in 1024+0 records out 1073741824 bytes (1.0GB) copied, 10.429498 seconds, 98.2MB/s USB Write Speed (Corsair Flash Voyager GT - CoreELEC) CoreELEC:~ # dd if=/dev/zero of=test.tmp bs=1M count=1024 1024+0 records in 1024+0 records out 1073741824 bytes (1.0GB) copied, 40.401781 seconds, 25.3MB/s Why is Corsair 4x slower?

[boxr4k]

It seems pv is not available on CE-NG so remove it from the command line like this :

dd if=/dev/sda of=/dev/null count=6291456

Since we don’t have the pipe anymore, we must specify where to send the data read by dd which is still /dev/null

pv measures the flow passing through it in real time and without it, you’ll have to wait until dd finishes to get the average transfer speed statistics. Remove count=### to read the whole device but it provides little benefit as there’s almost no variation between the start and the end of a USB flash drive like on rotating HDDs

[goober]

At the end of the day, all this didn’t even matter. I spent an hour trying to boot off the SanDisk and never got it working. No matter what I did to try to get it going, it just kept boot looping.

So I guess I prefer a slower drive that works than a faster one that does nothing!

[MauriceW67]

Does anybody know if it’s still possible to move CE’s data to EMMC if the Homatics is updated to ATV 14? I updated to the new 14.8.7475 today, but when I now run the “mount /dev/userdata /media/data“ command, I get an error:

mount: /var/media/data: wrong fs type, bad option, bad superblock on /dev/userdata, missing codepage or helper program, or other error.

EDIT: Using CE-NG 21.3 because of DV FEL support.

[goober]

My understanding is that with ATV 14, you can’t move it to EMMC.

[p750mmx]

Correct info. Android TV uses a different EMMC format on the SEI804’s (like the Homatics)