Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Booting Linux in One Second [pdf]

elinux.org

Post reply on HN
Print view View original

31–40 of 110 posts

Re: Booting Linux in One Second [pdf]

#31

Post by cm2187 »

psi-squared wrote: post #11
amelius wrote: post #3

An OS is recomputing the same functions over and over again when booting, so why not cache all of this, and make it instant?

How much do you cache? Hardware is difficult because you could add/remove/swap it at any time, and software information is already partly cached depending on your OS. IIRC, Windows a) builds a list of files to pre-load from disk based on your usage, and b) doesn't actually fully shut down by default (it goes to hibernate mode: http://blogs.msdn.com/b/olivnie/archive/2012/12/14/windows-8... ). I don't know if any Linu…

But an embedded system wouldn't change between two reboots.

Re: Booting Linux in One Second [pdf]

#32

Post by eitland »

cpach wrote: post #29
i336_ wrote: post #8

Earlier quoted context omitted.

I once had a brief conversation (via Reddit) with a guy using JACK for a commercial (embedded Linux) networked audio system. I wasn't aware of the exact specifics but I understand it was Sonos-style, embedded boards inside speakers. He wasn't particularly interested in talking about the work his team had done :P I think that proves your point.

This is a quite paradoxal phenomen that happens to some open source projects. The code is open and free, but some of the experts hold their knowledge very close to their chest. It's totally understandable since they of course need to make a living somehow, but it kind of defeats some of the benefits with open source.

Roughly speaking it levels the playing field. With Open Source smart people and companies that attract smart people get an edge.

Re: Booting Linux in One Second [pdf]

#33

Post by vidarh »

bnegreve wrote: post #28
kalleboo wrote: post #23

Earlier quoted context omitted.

I wonder if there are any disk suspend systems that are smart enough to tie into the VM system... instead of a specific monolithic RAM-sized suspend file, just flush everything to virtual memory on suspend, and on awake only restore the minimum needed for the OS and accounting, then start swapping everything else in lazily as the user needs it (or in advance in the background if the user is idle)

Given that the disk is several orders of magnitude slower that the RAM, lazily re-loading pages from the disk to the memory can induce very visible and annoying loading delays everywhere (e.g. change tab in browser takes 3 sec). I'd rather wait 10 extra second than having a slow computer for the next 10 minutes.

Or you could go for the hybrid, of having 10 seconds of reduced performance while the system keeps aggressively loading pages.

Re: Booting Linux in One Second [pdf]

#34

Post by cpach »

badpenny wrote: post #25
currysausage wrote: post #24

Earlier quoted context omitted.

With my Nexus 4, rebooting every other day is a pretty good idea. Usually fixes my Wi-Fi, 3G, and all sorts of weird bugs for the moment. Don't know how representative this is for Android phones in general, but most Nexus 4 owners I know tell me this is normal. Oh, usually once a week, sometimes twice a day, the phone will do the rebooting for me. Sometimes when it's in my pocket, sometimes when I try to watch a vide…

I used to have a Nexus 4 and had the same problem. I now have a Samsung Galaxy S3 and still have the same kinds of issues, just not as frequently. From what I've heard from other people I think it's fair to say it's a general problem with Android-based smart phones.

I reboot my iPhone 5s every two or four weeks or so. Sometimes that just the easiest way to deal with some bugs (e.g. Airplay issues). Since it happens so infrequently I don't really mind the boot times that much. But all else being equal, of course faster is better than slow :)

Re: Booting Linux in One Second [pdf]

#35

Post by cpach »

eitland wrote: post #32
cpach wrote: post #29

Earlier quoted context omitted.

This is a quite paradoxal phenomen that happens to some open source projects. The code is open and free, but some of the experts hold their knowledge very close to their chest. It's totally understandable since they of course need to make a living somehow, but it kind of defeats some of the benefits with open source.

Roughly speaking it levels the playing field. With Open Source smart people and companies that attract smart people get an edge.

Good point! But the hazard is that the we risk getting FOSS software that is unnecessarily complex to configure because of this.

Re: Booting Linux in One Second [pdf]

#38

Post by fulafel »

kalleboo wrote: post #23
akie wrote: post #17

Earlier quoted context omitted.

After which you'd have to load 8Gb from disk/SSD in under 1 second...

I wonder if there are any disk suspend systems that are smart enough to tie into the VM system... instead of a specific monolithic RAM-sized suspend file, just flush everything to virtual memory on suspend, and on awake only restore the minimum needed for the OS and accounting, then start swapping everything else in lazily as the user needs it (or in advance in the background if the user is idle)

Linux does approximately this.

Re: Booting Linux in One Second [pdf]

#39

Post by sharpneli »

legulere wrote: post #14

1 seconds seems a rather arbitrary measure. In HCI 100ms seems to be the value that's generally taken under which an action seems to be happening instantaneously [1] It's quite sad to see how we hardly use the computing power offered by our modern computers to actually speed up common tasks, but instead bloat it up with all kinds of stuff. [1] http://cogsci.stackexchange.com/questions/1664/what-is-the-t...

I completely agree with the general idea. Bloat is too high and has far reaching consequences.

And it's actually even worse than that. 100ms is nowhere near enough to be sufficient for all applications.

For audio applications you need 10ms [http://superpowered.com/androidaudiopathlatency] which incidentally is the reason why Android is unusable for any interactive audio applications.

Same applies to Virtual Reality. You need sub 20ms latency for head tracking to eyes latency or you will feel sick (effectively less than 10ms from the cpu side. Display scanning is summed on top).

The 100ms figure applies mostly to traditional UI workflow. Whenever you need a realtime interaction that's 10x too high.

Post reply on HN