Slashdot Mirror
No More Rebooting?
blankmange writes: "This headline caught my eye: 'The End of Computer Rebooting.' Seems that there has been some new developments in memory technology: The new thin-film technology that could give rebooting the boot is based on resistor logic rather than the traditional transistor logic used in most PCs and other memory-enabled devices. It also is considerably faster than current memory systems and holds the promise of reducing the time required to transfer and download multimedia content and other massive files. This is great news, but what am I going to do with the extra hour or so a day?"
How else are we supposed to fix things when everything just stops working...
Rebooting is always a great way to fix things.... they even used it one of the star trek's once.
And how the hell is windows supposed to work?
Like your computer, you need downtime (sleep, walking the dog, eating, etc).
If you are an avid computer user, you may only get your downtime when your computer is rebooting. This is especially true in workplaces where people are "chained" to their computers trying to finish a project, etc.
Those ergonomics posters on the wall do very little to get an average 'puter user to take care of themselves.. reboots served some of this purpose.
(Maybe that is why windows crashes so much - it's Bill Gates' gift to the employee!)
In any case, perhaps all offices should institute a staggered mandatory 15 minute inactivity period every couple of hours for each active computer.
Goat sex free since 2001
The press release doesn't really go into detail, so I don't know how similar (or disparate) the respective IBM and Samsung solutions are. They do both have the same net effect for users: non-volatile main memory.
This is cool stuff, but what hasn't been said is that as long as operating systems and applications leak memory, there will be a need for reboots.
Ciao.
on your DOS 286 did you ever try to load any TSR's or drivers for say sound or yes a network card (ok you got me it was a 386.) Hell have you ever booted a dos6.2 dos recovery disk it takes forever to load mouse and cd drivers it takes just as long for me to load WinXP on my laptop as any other os. If you are about to say strip it down yea it loads great if I load dos 6.2 w/ no drivers or anything I cant do much.. same thing w/ XP now just because this is slashdot I need to mention linux so if you have a stripped down version of linux it will load very fast unlike redhat (also had to mention redhat) which load very slow unless you tweak it, now that striped down version of linux you are running will run very well for what you tweaked it for but not for other general stuff.... That's just my opinion I could be wrong
This must be Thursday, I never could get the hang of Thursdays.
Back at the dawn of time, I was programming a (Data General) Nova II mini-computer which had "core" memory (which is where the term "core dump" comes from). Core consisted of tiny doughnut (ummm doughnuts) shaped magnets with (read/write) wires through it. It was incredibly slow by today's standards, but it did retain memory even when powered down. I'd shut the machine down at the end of the day. The next morning, I'd turn it on, and immediately pick up where I left off.
[Insert pithy quote here]
Hmm. Rebooting nowadays with 'traditional' OSes is to flush inappropriate state information out of the memory - an unusual sequence of events resulting in the system getting into a state it should never be in during regular operation....this might be either accidental (a crash) or semi-deliberate (an upgrade of a software component which needs a reboot to get it co-ordinated with the rest of the system). Having memory which maintains this state information will make the problem worse, not better!
What's needed here to achieve systems that don't need rebooting is operating systems which deal with all of these unusual events and states correctly..this means they'll catch errors and will be specifically designed to allow things like dynamic update to system compoents. I'm probably a bit biased but the best example a no-more-reboots kind of environment I see today is the OSGi.
True, you could pull different values (aka 1, 0, hell 2, 3, 4, 5) if you want) from resistance values.
I guess never having worked with resitance values as absolute, It never crossed my mind good point.
I think (I may be wrong) these magnatite are almost like "core memory" from days long gone, where a single bit was represented in memory by a single wound ferrite core, and I guess like you said pull distinct values from diffferent residtance values.
This sounds, upone rereading the article, even more like core memory on a film.
My dad saw, and showed me picture, yes im that old, of core memory being produced at IBM , thousands of ferrite cores beinghand tested and added to boards. Not bad in concept , but then again bubble memory for those old enough to remeber was a great concept too, there is still one Bubble manufacturer out there I am aware of upwards of 2 gigs, theyre using it as a solid state drive instead of ram (non-volitaile with somethinglike a 500? G shoch rating ?!?!)
Sig went tro...aahemmm.....fishing........
It seems to me that we're not understanding how this could be set up.. Why not have this as a device in your machine, that has an interface to the BIOS, where the user can set/format the unit to fool an OS into treating this nv memory-space as a fat32 or ext2 disk? You go into the BIOS, flush out the nvramdisk and 'reformat' and you are ready to re-install your OS should that become necessary. The rest of the time it runs as a really fast 'C:'... There's no need to replace normal RAM as your actual main memory during operation. Windows need not be aware of what is actually happening, you just boot/use faster, that's all.
Uh, have you seen WinXP's hibernate feature? On my 256Mb Athlon desktop, it writes the RAM to disk and shuts down in under five seconds, and comes back up (from wakeup keypress, through POST, then writes disk to ram) and is fully usable in twelve seconds. I've hibernated it with dozens of running processes and services, and not yet seen any problems on restore. I even took it down and brought it back up during a game of Deus Ex, and just kept right on playing where I'd left off.
Given a reasonably reliable OS, you should only be wiping the RAM when the system changes significantly, e.g. switching kernels or hardware. XP's hibernate feature demonstrates that merely turning the power off shouldn't require you to shut anything down. Unfortunately, I've yet to see anything that works as well on my linux boxen, including my laptop. Suggestions gratefully received!
If you were blocking sigs, you wouldn't have to read this.
I can't stand these Instant-On PCs that don't actually turn off when shut down. There are many times when you just want a clean start, without having the (possibly flakey) state of the system retained across a shutdown/startup cycle. With these "convenient" systems you have to pull the power cord to actually turn off your machine.
The real innovation would be for mainstream OSs (read: Windows) to be stable enough that they don't need to be restarted often. The time it takes to reboot your machine is not an irritant if you hardly ever have to do it.
The article wasn't any better written than the summary. It seems like this is suitable as a replacement for flash memory, not for either disk (which is huge), or for RAM (which is really fast). Of course, having a flash-like technology be cost-effective would change things; you could keep a copy of system memory as it is when it has just been booted (but before it initializes devices) there. Then you "reboot" by copying the virtual memory table from the nvram to main memory, and the system is immediately ready to initialize devices and run.
It would also be useful if programs could put some of their data in the nvram region, so (for instance), your emacs buffers don't go away when the power goes out. It would also be a good place to put write buffers, such that, as soon as the data is written to nvram, it will definitely make it into the filesystem, whether or not you lose power. This means that you can accumulate more dirty buffers safely and write them out in larger chunks, which is more efficient.
Keeping everything in nvram (if that were fast enough) may or may not be a good idea. You'd still want to reboot on occasion to refresh the system (load a new kernel, e.g.), but there's no particular reason you'd want to reboot at exactly those times when you power down and back up. Of course, you'd need everything to be hotswappable (replace the processor with programs running?) and restartable (disks have to be told to spin up, e.g.).
I just assume that electricity costs about 10 cents per kilowatthour. That's pretty close to what it is here, disregarding your disbelief. It costs a similar amount of money to install solar panels and get "FREE" electricity. But I've heard that it takes as much electricity to make a solar panel as all the electricity you'll ever get out of it in its useful life.