Firmware Update Kindle firmware 5.12.1

[Sirtel]

Originally Posted by patrickt

I suspect the Oasis line will die as the Voyage line has done.

There will always be a premium line and it'll probably include... you guessed it, buttons! Even Kobo has seen the light in this regard. And B&N with their Nook. Enough people want buttons and larger screens so that those are not going to disappear.

[j.p.s]

Originally Posted by NullNix

What is "turning completely off"? I'm fairly sure that the sleep mode Oasises go into halts the CPU, but the RAM is still powered because a full boot cycle takes about two minutes and nobody would want to wait *that* long.

CPU halt with RAM refresh left on describes the sleep mode of all kindles until the last few years. The newer kindles do that for about 30 minutes, then go into a state that is hibernation or very like it (save RAM and context to disk then power off).

[NullNix]

Originally Posted by JSWolf

Completely turned off is when it's not in sleep mode. When it's not slowly draining the battery while it's not used.

No device with a lithium-ion battery has a state in which the battery does not slowly drain when it is not used, both from self-discharge and because the battery circuit is required by contract to always be powered or they won't license you the patents to let you build one.

But more generally... modern machines' power architecture is far more complex than you seem to think. They have a wide variety of power states in which different components on the motherboard are at different power levels: the machine typically knows the levels available, the clock frequency (if applicable) each level implies and any appropriate dividers, the amount of power they save, the dependency graph between the various power-variable components (so you don't turn something off that something else that hasn't been turned off needs to have powered on) and the state transition cost (both in power and time). There is no single 'sleep mode'. There is a ferociously complicated and endlessly oscillating mesh of changing power states. With appropriate debugging tools, it can be hypnotic to watch the state transitions, particularly if some components are being reawoken on overlapping but distinct regular schedules: you get moire patterns and everything.

It is likely that significant parts of the system are unpowered even when you think the thing is in use (the most obvious part being much of the e-ink display). It is also likely that significant parts are powered even when you think it is "turned off". You can't tell just by looking at it, and you certainly can't tell by looking at the incredibly unreliable meter which is battery percentage (a proxy for battery voltage). The operating system on an embedded platform like this one will have a good idea, but it might not be too willing to tell you. (On larger systems, even the operating system is probably too far up the stack to tell just what power state various parts of the system are in.)

The question is not "is it turned completely off?". The question is "*which components* are powered down?". Unfortunately the RAM is one of the most power-hungry non-solid-state classes of components out there, and also one of the hardest to turn off or even turn down. I vaguely wonder if sleep mode saves some power here by shuffling everything into one RAM bank and turning off the other one to save power, but given the degree of change this would require to the Linux kernel, I really, really doubt it ("ferociously hard" doesn't even *begin* to describe it). It just seems like it would be a nice thing to do (but they're almost certainly not doing it).

(If the Sony isn't running Linux, it's possible it was running an embedded OS that allowed faster boot and/or possibly a RAM compacting scheme that allowed it to shuffle literally *everything* out of much of the RAM and shut it down. But I have even less idea if that might be true.)

[NullNix]

Originally Posted by j.p.s

CPU halt with RAM refresh left on describes the sleep mode of all kindles until the last few years. The newer kindles do that for about 30 minutes, then go into a state that is hibernation or very like it (save RAM and context to disk then power off).

Wow, flash is good enough now that you can swap hundreds of megs to a relatively small bit of flash just because someone left the thing turned off for half an hour, and keep it up for years?

[rashkae]

Originally Posted by NullNix

Wow, flash is good enough now that you can swap hundreds of megs to a relatively small bit of flash just because someone left the thing turned off for half an hour, and keep it up for years?

I don't know how Kinde implemented this, but SLC has rated life of over 100,000 cycles, even over 1,000,000 was being advertised back in 2008. a 1 or 2 GB SLC cache would more than cover what Kindle needs to suspend for years.

[j.p.s]

Originally Posted by NullNix

Wow, flash is good enough now that you can swap hundreds of megs to a relatively small bit of flash just because someone left the thing turned off for half an hour, and keep it up for years?

Someone in power must have decided it was better than waiting 2 minutes for a boot.

Didn't you recently post that defragging modern flash memory doesn't cause enough wear out to be of concern?

Regarding your reply to jswolf, many ereader brands allow "complete" power down (boot required). I for one am annoyed by kindles not allowing power off. I am fine with sleep by default, but I want the option to explicitly shut the kindle down if I know I won't be using it for a while. All the theory you wrote about all the reasons lithium ion cells run down on their own aside, my non amazon readers are much easier to maintain when not in use than my kindles.

I haven't had trouble with my 7" Oasis yet. Only time will tell if its flash dies before something else. It has no option to disable deep sleep but newer kindles do. Maybe that says something.

[NiLuJe]

You'll notice that the brands that "offer" (it's often fairly non-obvious to actually achieve) a full power down happen to be ones that ship devices that don't take 7 minutes to boot .

[j.p.s]

Originally Posted by rashkae

I don't know how Kinde implemented this, but SLC has rated life of over 100,000 cycles, even over 1,000,000 was being advertised back in 2008. a 1 or 2 GB SLC cache would more than cover what Kindle needs to suspend for years.

Sure, but would amazon spring for a separate hibernation drive, much less SLC for any of the flash?

[NiLuJe]

They do have a slightly funky, possibly slightly custom storage controller thingy (look for falcon in the kernel sources).

[j.p.s]

Originally Posted by NiLuJe

You'll notice that the brands that "offer" (it's often fairly non-obvious to actually achieve) a full power down happen to be ones that ship devices that don't take 7 minutes to boot .

Yeah, that is a long time to wait, but if I'm not going to use it for a few weeks or months, it is a trade I'm willing to make, and it is very annoying not to have the option.

I did not like that the PB360 has no sleep option at all, but the iriver story hd, onyx m92, and kobo handle it well enough.

The sony dpt-s1 stays "on" for some generously long time, then "sleeps" for some generously long time, then powers "off". It would be nice to be able to adjust those times, but the values sony chose work well for me.

[ilovejedd]

Originally Posted by rashkae

I don't know how Kinde implemented this, but SLC has rated life of over 100,000 cycles, even over 1,000,000 was being advertised back in 2008. a 1 or 2 GB SLC cache would more than cover what Kindle needs to suspend for years.

Alas, (2-bit) MLC NAND brought P/E cycles down to 10,000 then 5,000 then 3,000.

Planar TLC (3-bit) was down to 500-1,000 until they were able to mitigate with 3D TLC going back up to 3,000.

Unfortunately, QLC (4-bit) further reduces P/E cycles (I think it's back down to 1,000).

I doubt deep sleep works same as hibernate on PCs where RAM is offloaded to flash. The flash used in Kindles just isn't fast enough for that. They're likely using TLC at best. No way they're using expensive 2GB SLC just for cache when storage on the base Kindles is just 4-8GB.

[j.p.s]

Originally Posted by NiLuJe

They do have a slightly funky, possibly slightly custom storage controller thingy (look for falcon in the kernel sources).

Originally Posted by ilovejedd

Alas, (2-bit) MLC NAND brought P/E cycles down to 10,000 then 5,000 then 3,000.

Planar TLC (3-bit) was down to 500-1,000 until they were able to mitigate with 3D TLC going back up to 3,000.

Unfortunately, QLC (4-bit) further reduces P/E cycles (I think it's back down to 1,000).

I doubt deep sleep works same as hibernate on PCs where RAM is offloaded to flash. The flash used in Kindles just isn't fast enough for that. They're likely using TLC at best. No way they're using expensive 2GB SLC just for cache when storage on the base Kindles is just 4-8GB.

Weren't you the first to hypothesize hibernation as the deep sleep mechanism? If not hibernate, then what?

In any case, the write doesn't need to be fast since no one is looking. Also, they probably have figured out how to minimize the amount of RAM that needs to be saved. As a guess, they could close any open document before save and reopen it as part of waking up.

The slightly funky, possibly slightly custom storage controller thingy NiLuJe mentions sounds intriguing.

[ilovejedd]

Originally Posted by j.p.s

Weren't you the first to hypothesize hibernation as the deep sleep mechanism? If not hibernate, then what?

In any case, the write doesn't need to be fast since no one is looking. Also, they probably have figured out how to minimize the amount of RAM that needs to be saved. As a guess, they could close any open document before save and reopen it as part of waking up.

Probably posted while the synapses weren’t firing up properly.

Besides, that hypothesis was made back when I only had descriptions of the Oasis to go by. After I got the PW4 and had actual experience with deep sleep and NAND performance, I highly doubt it works like hibernate on PCs. The PW4 had way more battery drain during deep sleep on, wifi on than with deep sleep off, wifi off.

SKIP_UPDATE_CHECK has mostly solved my battery woes when leaving wifi on but honestly, I just don’t notice any difference with deep sleep (apart from waking up taking longer).

[93terp]

I'm not 100% certain that this is related to the latest firmware update, but on my new Kindle Paperwhite (2018 version, WiFi + 4G LTE, 32GB) when sideloading files from Calibre, after ejecting the device, there's a new dialogue box that says the following -

Transferring Files To Kindle
If you transferred files to this Kindle, battery like will be impacted while the device processes these files so you can search them. For best results, charge your device after transferring files.

What's curious is that both of my Voyages are also running 5.12.1 but neither of them have this dialogue box after sideloading files from Calibre.

[Quoth]

It's true though. So keeping connected to charge after ejecting if you copy a lot of books in one go is a good idea.

I've never seen that message.
It won't be due to calibre but due to transfer and actually could apply to lots of files in one go via USB, 4G or WiFi. They have to be indexed and maybe thumbnails generated?