Slashdot Mirror
MRAM in 2004?
amberspry writes "As previously reported here and here. Wired has yet another update on MRAM here. They give hope by mid-2004 we will see devices with faster boot up times and using less power as a 'vastly accelerated timetable is being implemented.' Gotta love joint ventures."
Yummy!!!!
If only this would come out sooner... IMagine being able to turn on your computer and not having to wait to see it load!!! Not even come out of hibernation! Amazing! Bring to the UK now!
I'm sure there probably isn't anything to worry about, but isn't there a chance of problems if you put magnetic things near storage media?
Try not. Do or do not, there is no try.
-- Dr. Spock, stardate 2822-3.
comes in multiples of 2 ternary digits (2 tits).
I'm glad to see that the hardware industry is producing vaporware now and that vaporware is not exclusive to the software industry.
BTW didn't Bill Gates promise instant booting PCs five or so years ago? My new machine takes a full two minutes to boot.
So I can keep all my expensive pir- er, rightfully purchased Photoshop clones open at once! No... 8gb wasn't enough...
How does MRAM fair against magnets?
But when will holographic storage arrive on the scene?
Even though they calim it will be cheaper to produce and have higher desities, you can be sure that it will cost an amr and a leg for the frist few years. I don't expect this in mainstream systems for quite a while. Still, really sweet looking tech.
"A coward dies a thousand deaths, the brave but one."
They're loud, slow, and ugly. The sooner they go away, the better.
-Libertarian secular transhumanist
"...simply reach out and touch an on/off button to turn off Windows in lieu of going through a ritualized shut-down procedure."
who says we will be running windows by then?
I hope not....
Matrox used WRAM.
I can't wait to be called out to degaus someone's ram after their system crashes.
I can't wait till this technology can permanantly remember data. AND it gets cheap enough to replace the spinning hard drive. Speeding up the memory read/write times and reducing the memory bottleneck could effect your pc much more than upgrading from a 1.8 ghz to a 2.0 ghz processor.
Slashdot Syndrome: the sudden, extreme urge to correct someone in order to validate one's self.
MRAM @ whatis.techtarget.com
MRAM @ Webopedia
MRAM @ German Wikipedia
Well, geez. If I kept the SPARC on all day I think it'd melt right through the table. To say nothing of the fact that I couldn't hold a conversation without yelling.
Not that I can, anyway, actually.
This is damn sexy technology. Almost makes up for the vaporware that was Keele Memory systems. And if I hear one more whiny person say, "no, quantum computers are coming in two years! Have patience!" I think I'll go destroy something expensive...
That was actually pretty good for slashdot.
A little more indepth view of MRAM can be read here.
Does anyone know if MRAM will be sensative to external magnets? Aka if I bump my portable mp3/ogg player into a giant fridge mag will I lost my data?
Apple free since 1990!
I timed BeOS on my new machine. It gets from power on to fully usable desktop in just under 16 seconds. Too bad the only thing it's good for anymore is booting fast :-(
Most people will just grab a beverage or something during the minute (or less) it takes most PCs to startup. I would think most of the people who keep their PCs on 24/7 do it for P2P or [Seti|Folding]@home or possibly to prevent wear and tear on the hard drive (spinning up the hard drive wears it down faster than anything).
Slashdotter are stupid and biased.
2004? Isn't this the same year nanRAM was supposed to show up? I'd rather ditch anything mechanical or magnetical in favour of stuff that doesn't deteriorate or move.
Hate me!
Is this the way to large, cheap, solid state storage devices?
Is the HDD on its way out?
Or is it just something else to compete with SD, CompactFlash, MemoryStick, etc? The article talks about cell phones and PDAs which makes me believe so.
Technically, it sounds like a miniature magnetic drum.
I don't need no instructions to know how to rock!!!!
This technology reads like it is based on the old
Magnetic Bubble technology. Late 70s / early 80s.
Except Bubble memory cards were about the size of an IDE Drive.
This memory get's rid of the need to save your settings to the hard disk as you power down. But when your computer dies, you don't want the "bad" settings saved to the hard disk.
It will be interesting to see the new breeds of virus that this brings out.
Does this mean that when my cell phone rings, my speakers AND my RAM are going to go nuts?
Will my pc run faster if it is facing polar north?
Celebrate Steak and a Blowjob Day!
This would be a great thing for power bills as well.
Lots of times you want to keep a machine up all the time, like in my case when it's serving up a webpage or two and acting as a print server. But I'm sure there are also plenty of people who leave their machines on all the time just to avoid the startup/shutdown time. I know I do it with my laptop just to avoid the un-hibernation.
With power supplies averaging, oh, 300 or so watts, that can mean decent savings when you figure it running 24x7.
Could this someday replace the hard drive? I realize that hard drives rely on magnetic data storage, but it's done in a different way than MRAM. Perhaps when MRAM technology can hold more, we could possibly do away with the hard drive and have even faster loading times to your programs.
I wonder how MRAM would handle memory errors though.
...your computer will erase your credit cards.
Clearly, my computer will startup no faster than it does when coming out of Standby mode (which stores the state of my computer in RAM, but requires that the PC remain plugged in). So, what do I gain? Basically, we get Standby mode that works even when you unplug the computer. And, that's still no improvement to the "startup time".
So, who needs their cell phone or PDA to startup faster? Most of these devices are pulling straight from some flavor of RAM during startup, already.
How often do you reset your iPaq? Just when it crashes, and it only takes 5 seconds, anyhow.
What about that annoying startup time on your cell phone? Let's see, only when the battery falls out do I ever exercise that feature.
If MRAM is really 6 times faster than today's static RAM, that's wonderful, but it will have little impact on startup times (see Hard Drive I/O-blocking).
Here's a better link for more info on MRAM. Pretty graphic of an MRAM cell.
Any technology distinguishable from magic is insufficiently advanced. - Geek's corollary to Clarke's law
I *WANT* memory that clears itself when you turn off the computer.
Otherwise, What good will it do to reboot after Windoze Blue Screen(tm)'s on me?
I'm just sitting here watching the wheels go round and round.
If you open your mind too wide, people will throw trash in it.
It will take at least few years before density will be high enough to use MRAM in a PC.
RIGHT ON!
Just one big glob of memory, full of persistent objects. No more goofy file systems.
Except for the fact that due to all the memory leaks and other programming issues in Windows, you'll still need to do your daily hard reboot. This will just make it slightly faster.
--
Luck is just skill you didn't know you had.
You misspelled DEAN in the headline and it looks like you should have used 'smoking' instead of 'ventures' at the end of the article.
I just can't for that new memory does ... does, er.. hmm, what does it do again?
"A door is what a dog is perpetually on the wrong side of" - Ogden Nash
is now 60 seconds of pr0n viewing. Viewing time that would have been lost to oblivion. Thank The Maker.
Shop smart, Shop S-Mart.
I don't want to sound too cynical here, but I just can't seem to get so excited about Motorola working on new innovative technology and continuing on with it. I remember when Motorola phones were the way to go. Even more dramatic an example though is the whole PPC chip. There was once a time the chips they produced for the Macs were just slightly slower than Intel's chips (in terms of MHz...but we all know that doesn't really matter for true performance). But then they seemed to take naps that lasted for years while AMD and Intel kept improving chip speed and performance. Sure...Motorola may be working on this now, but from what we've seen in the past, I wouldn't be surprised to see them resting on their laurels and letting the world pass them by yet again.
"He uses statistics as a drunken man uses lampposts...for support rather than illumination." - Andrew Lang
MRAM @ How Stuff Works
Mmmmm...RAM Sorry I couldn't resist.
Chaos reigns within.
Reflect, repent, and reboot.
Order shall return.
Strange, my computer crashes each time I hit the "degauss" button on my monitor ...
"A door is what a dog is perpetually on the wrong side of" - Ogden Nash
people...get a clue!
with mram you perform one good clean boot...initializing everything...then "lock" that into your mram as the default start state...OS loaded and ready to go...then every boot has that "clean" experience by zapping to the saved state...if you're system needs a "boot" to get back to clean you just zap there instantly instead of waiting. Its not about needing to reload slowly from drive to get to clean state...clean state is always available. Bring it on!
Unlike conventional high-speed memory devices, MRAM uses magnetism instead of electrical charges to store data -- making it, in a sense, a back-to-the-future technology
:)
Does this mean that our 60 gigabyte hard drives will be replaced with 60 jigabyte hard drives?
You can't take the sky from me...
As previously reported here and here. Slashdot has yet another dupe article on MRAM here.
Wow... this guys even SAID this is a dupe and he still got his story posted. Sweet!
Karma: NaN
Karma: NaN
Is Rambus involved? I didn't see any mention of Rambus, but someone might want to check... your backside for a knife... er... I mean check the US Patent Office to see if they've patented the IP.
-- No sig for you!
We have ASSRAM right now. It's not very fast, butt it's quite powerful. Speeds run from 1-3 FPCS (Fudgepacking Cycles Per Second).
How much fudge can a fudgepack pack if a fudgepack could pack fudge?
Comment removed based on user account deletion
Bubble memory was serial (internally). A rotating field moved the stream thru bubble detectors and generators. It was similar to mercury delay lines with bubbles instead of sound pulses.
This tech looks like it's actual RAM.
Like access times and whatnot?
The article talks about PDAs and cell phones, which doesn't blow my mind compared to flashram that we have now, except maybe 5% more battery life.
How does this stuff stack up against something like DDR? Can it come anywhere close to keeping a P4 with an 800mhz FSB full of data? Will it replace my system RAM, HDD, both, or neither?
I don't need no instructions to know how to rock!!!!
The story is almost worthless marketing drivel. How about answers to some very basic questions like:
What is the capacity?
What is "extremely dense" in quantitative terms, and how do they achieve it?
If it's really going to be a "universal RAM replacement", how does it compare with the 512 Mb DRAMs recently announced?
There are many more similar questions, but answers to these three would be a start.
I can see it now, the hordes of inept that were suckered into forwarding e-mails in hopes of obtaining $10 for each e-mail (because it was being traced by the new e-mail tracing thing!), and deleting that obscure Windows system utility because it was a virus, can now be suckered into purging their computers of evil bugs by simply "resetting" the MRAM by holding large magnets to the side of their computers!
This post was brought to you by the number 584811 and the characters / and .
Though I think technology like this will really make a difference because there are only so many practical applications for speeds of 3Ghz, yet there are still issues in a computer's speed such as loading time that don't usually get much attention
So now what would the next step be?
I for one am wondering when we'll reach the instant gratification of Star Trek computers, but that's just 'cause I'm a geek that way :P
Kick in the Head
Found on someone's web page of sig files....
"The Motorola 6800 had an undocumented assembly opcode that earned the
mnemonic 'Halt and Catch Fire'. It was used by the factory to test the
address bus. It's harmless when the chip is hooked up to a test stand or
normal RAM, but hook it up to core memory and it really would fry."
- Unknown
Anyone remember that Episode of Doctor Who when he took the core memory through time? Normal RAM wouldn't work. Core memory has a state and therefore could make it through time.
"Technically, it sounds like a miniature magnetic drum."
This is more like core memory. Plus this is faster than static RAM. Which itself is faster than DRAM.
BTW I wonder what this means for OS design? No more caching, and other techniques for dealing with a slow device.
Linux is not immune from memory leaks (or any other software problems for that matter).
Why would someone turn 'off' their PC if they don't turn off their TV?
Leave the machine (in my case, a Mac) in standby, just like a TV. Touch a key, the power button, or the remote, and it switches on in less than 2 seconds...
From all external appearances, my Mac is off, except for the glowing power button.
GPL Deconstructed
Actually that would be pretty good for power failures...
No more work lost under Office when my roommate switch her hair dryer on...
Iraq: war to save the U
With that said, there's no consumer application at the moment other than video editing that comes close to using the bandwidth of modern IDE for sustained periods, so I imagine this tech could popularize centralized network storage for consumers. the appeal of a wireless tablet PC would be pretty strong if it didn't have to waste weight, money and batteries spinning platters...
Good thing the last two generations of game consoles have gotten us inured to load times! If there is such a moved toward consumer-space network storage, I suspect MRAM will mean faster OS (and integrated application) loading, but significantly slower load times for infrequently-used or large third-party apps.
good going ! less power use = less heat, less heat = less fans, less fans = less noice !! great !!
Please arrange your letters in the correct order.
End Of Message.
Unix had this 30 years ago.
There are different usage models, you know, without calling someone out on FUD.
My Mac, for example... on 24/7, goes to sleep after three hours.
Goes from something like 120W during average use, to 80W with monitor in sleep, and then to 9W when the whole system goes to sleep. Possibly less.
However, one thing that I don't have to deal with every morning is loading an OS, loading apps into memory, and the extra disk access as part of that process, since it's all already in memory.
So in one compromise situation, my drives aren't spinning up every morning, they spin down after 20 minutes if inactivity, and my monitor, video card, CPU, and eventually the rest of the system power off one by one until the whole thing is asleep...
GPL Deconstructed
"unpatriotic"
-The american government to anyone who dares criticise their actions
I'm struck by how much the HowStuffWorks picture of MRAM memory (*) looks like the donut-on-a-wire ferrite core memory. All that's missing are the 150-ohm terminating resistors.
I like the idea of a HD-less instant-on PC. One of the great things about my Palm Pilot is that the kids can turn it on and off without any "shutdown" process... although all my kids have known how to shut down Windows properly since they could understand the "To turn off press Start" concept.
On the other hand, it's already hard enough to restart a locked-up PC when the so-called power switch doesn't have anything to do with the power. How will I fix a PC when pulling the plug doesn't even reboot the OS?
Stressed? Me? Of course not. Stress is what a rubber band feels before it breaks, silly.
'Unlike conventional high-speed memory devices, MRAM uses magnetism instead of electrical charges to store data -- making it, in a sense, a back-to-the-future technology based on the same laws of physics that enabled the creation of audio and videotape recorders as well as hard drives.''
To say nothing of drums and original core memory!
Great minds think alike; fools seldom differ.
Once upon a time Motorola chips actually ran at higher clockspeeds than the fastest intel chips. For example, the PowerMac 9600/350 ran at 350mhz while the fastest Pentium II clocked in at only 300mhz. Reference.
They're loud, slow, and ugly. The sooner they go away, the better.
You forgot to add 'consume a lot of power and generate a lot of heat'.
How wil i be able change network settings when i'm unable to reboot?
1) The amount of heat these things generate is excessive. The heat sinks on the MRAM will be about the size of the heat sinks on a 3.2Ghz P4.
2) MRAM emits minute amounts of beta particle radiation, meaning that it actually has almost as high a risk of causing cancer as, say, smoking a cigarette once a week.
3) The real-life density limit of MRAM is about 1/4 that of modern DRAM. The switching hardware around it requires large amounts of energy (MRAM switching is actually much less efficient than DRAM).
4) The lifecycle of MRAM is considerably shorter. First generation MRAM will last about 2 years before the ferrite decomposes rendering it unusable.
All popular OSs have memory/resource leak problems. Even my OS X laptop gets s-l-o-w after a few days or weeks of use, and logging out and back in speeds it up again. I notice it most when I use anything from MS Office, for some reason...
Almost all of this can be laid at the feet of pervasive use of languages that require manual memory and resource management. Writing leak-free-C is apparently beyond most normal mortal programmers, or even the wizards who write things like Apple's Quartz layer or XFree86.
All these issues are going to be magnified with MRAM-based machines - if sleep/hibernate mode is free, it'll get used more, and more systems will have longer routine uptimes.
To a Lisp hacker, XML is S-expressions in drag.
"How many people keep their computer on 24 hours a day simply because they can't stand to sit around for four or five minutes waiting for it to boot up?" he asked. "
If you wait 4-5 minutes for your computer to boot up, what's the chance that you're going to spend the money to upgrade to a system using MRAM. My computer at home isn't what I would consider top of the line and it boots up much quicker than that.
You can actually buy FRAM still. (Ferro RAM). I'm not certain what the difference between this new "MRAM" and old FRAM. Maybe they just need a new marketing buzzword to differentiate a next generation product?
“Common sense is not so common.” — Voltaire
Does it mean memory leaks will linger forever? Microsoft better watch those memory leaks then...
For those wondering what the use would be of an instantly rebootable computer, they obvously haven't been on the phone with NTL support asking you to reboot your machine after every change.
Or they're not running mission critical servers were every minute of downtime costs thousands of units of whatever strong currency you're using.
But fast or not, it will not last. I mean, sure I can (could: I haven't tried) boot Windows 3.1 on my 1.4 GHz P4 in 3 seconds flat, but so what? Microsoft is going to always use 150% of your resources just to make sure they're never beaten on the feature list (otherwise known as peeing contest) and the insecure open source community is never far behind (but always so) with ever growing feature lists too as they want to catch up. (I DO love Linux, OK?)
So MRAM technolgy may be all that good, but it will be abused without a doubt as soon as MS get their greasy hands on it and fit all their development machines with it.
To be fair NTL tech support is OK, once you get to them (after about 2 hours) and once they start listening to you (about 30 to 45 minutes).
Does this MRAM come with the Hemi? Or is it just the MagRAM?
My old 8-bit Atari (1200XL, not that anyone asked) could "boot" (no DOS) in about two seconds.
"A great democracy must be progressive or it will soon cease to be a great democracy." --Theodore Roosevelt
Does this remind anyone but me of the ferroelectric memory cells of about a decade ago?
Smaller than DRAM cells, faster than SRAM and nonvolatile as well. They did actually make it out into the real world, several devices made today include a dozen or so F-RAM cells, but they certainly did not take over the world.
One thing that does shout "vaporware" to me is that the articles I can find are all really sparse on details.
Also, how compatible is this technology with common (or esoteric, for that matter) silicon technology? If it's not, can we use the same technology to build processors, etc.?
How soon do we actually get to see a 256 MBit MRAM device? How much will it cost in 2005? The answer to those questions will tell me a lot about whether this is enough to make people show interest in Motorola's stock again....
-- Ancient (IBM 1620 and Atari 400) Programmer
I timed my new home-built P4 system running XP the other day, and it took 30 seconds to boot to the desktop, including the BIOS POST. I don't think anyone waits 5 minutes for a PC to boot anymore.
If you are so busy that you can not wait for two minutes for your computer to boot, how do you have time to read slashdot?
If that two minutes is actually so critical to your day, just leave the computer running.
My biggest question is... I'm planning on a major upgrade to my system in the next few months for *cough*Half Life 2*coughcough* important personal reasons. Is this going to be cheap enough that I'd be better off waiting 6+ months instead of going with what's currently on the market? I'm certainly not going to drop everything just because new tech becomes available...
--Fesh
Kill -9 'em all, let root@localhost sort 'em out.
Texas Instruments had this technology years and years ago:
1) wafertape digital tape drive for CC40
2) some sort of solid-state wafer drive available for the 99/4A
All that plus 16-bits, hardware sprites, stereo sound, RS232, you name it.
Imagine the temperature difference in server rooms these will bring. As all of you know, heat is a big problem, and it costs lots of money to keep the temperature down. Or am I wrong? Actually how hot do these get?
1. Wake Up
2. Turn Computer On
3. Relieve Bladder
4. ???
5. Profit!
I guess everybody saw this coming...
I think mram should be used as an alternative, not a solid new type of ram that we "have" to use. we should have mram as a side function for faster filesaving, but not for system ram.we still need dram. and for those who are worried about the drm shit so a certain company doesnt lock your computer with their software only, just use a magnet ;) or do something that will clear the mram.
mram should be used a secondary. all it's gonna do if it's used as primary system ram is cause issues with broken operating systems, and if you get a virus? oh hoo, you're shit outta luck.
Maybe some technically familiar with this technology can answer this; Will an external magnetic field have any effects on devices using this technology?
As far as fast boot times go...one of Spider Robinson's characters had an interesting observation. He spent a thousand dollars getting a hard disk for his Mac (which is about right for the time the story was written) so he wouldn't have to wait for a boot floppy once a day. Yet, he spent 10 minutes every day fiddling with the water controls in the shower to get the temperature right. For far less than a thousand dollars, one could build a system to automatically adjust the water temperature in the shower, but we don't. Why does a slow booting computer bother us so much, but all these other things that waste more time don't?
That's nonsense! Sure, it's a well known fact that the current inrush is rough on the silicon. But, if you're reading this, you're a geek who will upgrade to the next hyperthreaded scsi-8 USB-7 quantum CPU as soon as Intel/AMD/TI releases it.
Now in light of our happy go lucky bombing of IRAQ and dependency on Foreign oil, we really should be shutting off those dogs that don't need to be barking.
may even, someday, allow us to simply reach out and touch an on/off button to turn off Windows in lieu of going through a ritualized shut-down procedure.
Hmm, if I touch my power button, everything closes out and I get a message saying "Windows is shutting down..." That technology isn't called MRAM. It's called ATX (power supplies, anyway) and ACPI.
slowly this is really annoying. You know that howoften it is posted, win98 doesnt come back and the nt kernel doesnt become unstable.
Just to make a point: my pc was rebooted 3 times the last 4 month.
Once because i downloaded some BS patches, once because a hd had a SMART error and had to removed and once because the new HD had to be build in.
Now, i have 4 or 5 days uptime (the hd came last week) and 57MB kernel memory. Doesnt seem too leaky, considering the PC is actually in use and not a "seti and nothing else" machine.
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
...but my WinXP machine has been up now for over three weeks straight. Doesn't quite compare to the Linux server in the corner, but there is nothing wrong with Windows itself. Now, if you run buggy programs that leak memory, I'm not so sure how well Windows deals with that. But I'd say that "insightful" complaint lost it's validity around Windows 2000. Maybe not for 99,99999% uptime servers, but any desktop used doesn't give a fuck if they have to reboot once every other week.
Kjella
Live today, because you never know what tomorrow brings
CF, SD, MMC, Memory Sticks... bah! I wanna see some 1GB MRAM sticks about SD size and less than CF prices! Seriously, I've been waiting for these to show up on the market since I first heard about them about what 2-3 years ago now?
But I'm not trying to bash the companies working on the project, I'm seriously glad they've been working on such a technology, I'm ready to see it happen!
There are some BIOSes [sic] that allow you to schedule a power on and a power off time. I don't know if my BIOS has such a feature, I never looked.
Well, why do people always log in as AC if they want to post shit?
1) They will create MUCH less heat than common RAM. They dont need capacitors (which discharge creating head and have to be refreshed, costing power).
2) (smoking crack?) I dont even have the slightest idea why they should emit beta radiation, but even if they did, beta radiation isnt very good at penetrating anything. even if some electron could escape the plastic casing of the Ram package, is would surely stopped by your case.
3) Comparisons with flash memory are far more useful considering the proposed usage of MRAM. Also stacking would be potential way to increase packing density, because there is no need for capacitor trenches.
4) What part of "first generation" did you not understand?
I try to imagine how humanity would progress with your attitude:
Eddison: Why bother creating a lightbulb, it will only break after a short time nobody will need it.
von Braun: The thing could explode and even if it works, it just falls down again. Why should i create a rocket?
Einstein: Everthing seems relative... But i guess that we can never make use of it because we would need complicated machinery and mathematics. Lets paint some pictures instead...
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
It has to be "non-volatile, inexpensive, fast and low-power" hehehe...yeah...and fire...hehe...and lots of chicks...hehehe...that'd be cool...and we want it now...yeah...hehehe.
You mean, like, I bring the papers, Tony brings the dope, and Shane brings the lighter?
another technology we really dont need. If Award, AMI and Phonix would clean up their BIOS code you could get those faster boot times. The Linux BIOS project has mobos bootin into a full OS in less than 10 seconds.
My karma is not a Chameleon.
Right now a database transaction must be written to disk before it commits. This is because a database must be durable, meaning it retains all committed transactions after an OS crash or power failure. Since disks are terribly slow, and since transactions must be written to disk, disk speed is the limiting factor in database performance.
Large databases get around this bottleneck by having thousands of disks writing in parallel. This increases the speed, since thousands of disk heads writing in parallel are thousands of times faster than a single disk writing alone. The multiple disks are used for speed only; most of the space is unused.
Thus, large OLTP databases require enormous disk arrays, with a huge number of disks writing in parallel, to make the performance of disk adequate. These huge disk arrays can cost millions of dollars.
MRAM is non-volatile. And MRAM is so vastly faster than disk, that a single MRAM drive could sustain a write bandwidth greater than 10,000 SCSI disks in parallel. Therefore, 2-3 MRAM drives could replace the huge multi-million dollar drive arrays and offer better performance. This would allow extremely busy OLTP databases to use a small number of commodity parts, rather than using huge million-dollar disk arrays.
Extremely fast databases would become a commodity item, rather than something costing millions of dollars. Since the new HP 4-way Itanium2 server offers >100,000 tpm/c (greater than the old 64-way E10000!), a single 4-way i2 box with a few MRAM drives could probably power the database for Ebay.
Right now, operating systems and databases are designed around a particular storage model. Storage is divided into several levels. At the lowest level is large, extremely slow, non-volatile disk for permanent storage. Above that is RAM, used for running programs and caching oft-used data from the slower disks. Above that is cache, used for caching oft-used data in slower RAM.
But now, we'll have MRAM, which is non-volatile like disks, but faster than cache. And with memory densities growing according to Moore's law, soon we'll have MRAM memories as large as disks traditionally were, allowing large non-volatile "disks" that are faster than cache SRAM.
This completely reverses many of the assumptions made in designing OSes and databases. Will there even be a separation between disk and RAM any more? What is the point in having separate MRAM disks and MRAM memory? If an MRAM disk is faster than cache, then why have caches at all? Why have "paging" operating systems, why have OSes that cache disk data, why have complicated algorithms to reduce disk access? Why have buffers, if writing to a non-volatile file is faster than writing to a buffer?
Having a single, large, cacheless MRAM bank would seem to be the most appropriate scheme, with perhaps a disk used (with a swapping scheme) if the MRAM bank isn't large enough. A filesystem would still be required, but it would be in RAM. (Filesystem design would be very different).
The appearance of large MRAM disks will cause many things to be reconsidered.
Gotta love joint ventures.
Cool. Does this sound to anyone else like it's going to result in a lot of smoke and a serious case of the munchies?
Blog,Twitter
I can tolerate losing some bits of info in an audio tape but can you afford CPU memory getting corrupt due to stray magnetic fields ?
Core memory is making a comeback? I knew I should have kept that Honeywell H316 with 32KB of memory.
Will they also introduce a new form of paper tape made of stronger, more durable materials?
So now when I accidentaly use a magnetized screw-driver to replace components, I'll end up f'in up the memory as well as the hard-drive!! Yeay!!
From PC Magazine's article on MRAM:
As for replacing hard drives, this is on a pretty distant timeline. An analyst in that same article said:
So, 10 to 20 years for DRAM means at least another decade beyond that for hard drives, if at all. Cool technology, but it just seems like another obvious step on the technology path.
The bottom line: I don't think anyone on Slashdot is very surprised to find out that a technology will be around in 20 to 30 years that will let us persistently store stuff on something quicker and better than hard drives.
Excellent point -- I can reboot my Palm with a paperclip in the Reset hole. Surely there would be something similar in an instant-on HDD-less system.
.sig:
Hey, I've got a reply to your current
Macromedia + Doubleclick + Webcams + Microphones = WTF?! [josef.org]
The message box you displayed includes "doubleclick.net" in its text because the Flash animation is being served from doubleclick.net. If it were being served from slashdot.org, the message would read: "Allow slashdot.org to access your camera and microphone?"
Someone could design a Flash program that interacts with the user via microphone and webcam instead of keyboard and mouse. There might be a valid reason for this -- maybe a company is developing an in-house Flash-based videoconferencing system, or something. But Macromedia wouldn't want to enable such access by default, hence the dialog to allow it or not.
The fact that it's been a year or two since I last poked around and saw this option tells me that the idea of interacting with a surfer's mic/vid hasn't taken off. That's probably a good thing, 'cause I don't think we want to ever have a situation where your kids turn on the option for their favorite online game without telling you. An entertainment portal that watches YOU is a little too "1984" for me.
By the way, I'm with you -- overuse of Flash animation makes me look for the "Back" button.
Stressed? Me? Of course not. Stress is what a rubber band feels before it breaks, silly.
unless there is a manual method to wipe the shit clean when you need to. That's all I need, some corrupted piece of shit software in RAM preventing my machine from booting.
Geek used to be a four letter word. Now it's a six-figure one.
...A hard drive, split into millions of pieces, each bit containing a miniature read/write head which is independantly addressable? It sounds just like a hard drive without the spin, and with an equal number of heads to bits. Sounds interesting, but if it's nonvolatile as they say, wouldn't external magnetic fields and normal magnetic decay mean that there is a good chance of a corrupt bootup after an extended vacation?
The good thing is that this may discourage people (ahem MAC cough) from placing their motherboard in a case with a CRT? GOD I HOPE SO.
SDRAM is power hungry during sleep mode ( a few mA) and has a slow sleep/wake-up sequence. This is not very nice for some devices like cellphones.
I think the most likely use we'll see for MRAM in the short term is having, say, 256kB of MRAM in a cellphone for running the cellphone engine and using SDRAM etc for the extended features.
Engineering is the art of compromise.
The title of that article is a bit misleading.
Remember that MRAM is not really new technology. It's a new implementation of something that already was used in a computer half a century ago: In 1953 ENIAC got its 100 words magnetic memory.
And the photo of that sputtering machine makes it look like some cool contraption from an old frankenstein movie... Need I say more?
--- Hindsight is 20/20, but walking backwards is not the answer.
MRAM may only be supported by Palladium(tm)-approved mobos with heaps of DRM. If that is the case, this could be the "killer app" that makes you think about ditching your current generation machine for a police state media-player.
If so, please remember me. I'll buy your FREE (as in liberty) computer from you.
because I have been enjoined by this Holy Office to abandon the false opinion which maintains that the Sun is the centre
Ok, MagRAM sounds all nice and stuff, but are we going to get scalped on the price like we did on SDRAM and earlier DDR modules?
What does the media type have to do with having a filesystem or big glob of memory stored on it? Oh, well, access speed I guess would help with your glob. Don't worry, you'll see it come eventually.
When you burn a cd does ambient light destroy the medium? No. Just like a strong enough light source (for example, the burning laser in your xDRW drive) could destroy or alter the data on it. There is not going to be a strong enough field to do this accidentally, the fields put off by electronics are too small to be of much concern. If there even was a problem, parity algorythms could be used the same way they are in larger magnetic media (RAID). It is certain that a stong enough magnetic field could depolarize the data but I doubt the field is that "loose". Furthermore a magnetic field is inversly proportional to the square of the distance to the source ( E~1/[R^2] ) which decays rather rapidly as the distance increases. Electronic devices may operate at a high frequency but the power (amplitude)is too low to low to generate the field that would be required. Electic/Magnetic Field Equations
If you want to save money on power bills and want to avoid un-hiberation, get a Mac ^^
My Mac averages about 120W consumed in use, but when the monitor goes to sleep that drops to 50W, and when the drive spins down it's 30W, and when everything is asleep except for the RAM, it's at 10W...
And it'll still act like a webserver and printserver at 30W ^^
And waking up from sleep only takes 2 seconds. Who needs MRAM 3 years from now when you can have a Mac today?
GPL Deconstructed
"How many people keep their computer on 24 hours a day simply because they can't stand to sit around for four or five minutes waiting for it to boot up?" he asked. "I don't think anyone has researched that particular issue, but I'll bet there are a lot of them.
I guess it must just be a PC thing. I have a PowerMac, an iMac, two PowerBooks and my wife has an iBook. I can't remember the last time I booted a single one of those devices. On the laptops, you just close the cover and they go to sleep. On the desktops, they are set to sleep after a period of inactivity. And with Macs, going to sleep / waking from sleep is essentially instantanious....
So my question is, "Do you really need MRAM to solve that problem, or maybe just some lessons from Cupertino?
It is your personal duty to fight for what is right on a daily basis. Ignoring injustice is identical to approving
Correct. The magnetic field as you know it is created entirely by the movement of charge. Whenever you have a current (stream of electrons = moving charge), you will have a magnetic field.
As an interesting side note, while you might think electricity and magnetism are different, they are actually one and the same. Magnetism can be shown to be a direct consequence of Charge invariance (an electron has the same "charge" in any simply moving reference frame) and special relativity.
I remembered reading about magnetic ram 5 years ago in Scientific American. Cool it's finally coming (soon enough) to a store near us.
;)
Next item on my wish list is giant, cheap, space-efficient, superior performing OLED or similar display technologies. Hope that comes to a store near me as well
VIVA1023.com | Political Fashion.
"How many people keep their computer on 24 hours a day simply because they can't stand to sit around for four or five minutes waiting for it to boot up?"
Huh? I leave it turned on because it increases the machine's life, and because there is no reason to turn it off!
Will my flying car come with computers based on MRAM?
Stick Men
"The reality of it is that faster OS booting won't happen until, among other things, Microsoft adds save/restore peripheral functionality to drivers (making them even more complex and unreliable...)."
Alternatively, give a Mac a try. One thing which impressed the hell out of me when I first got my iBook, once I noticed that it was happening at all, is that closing the screen puts the machine to sleep, instantly. Open it up again and it's ready to go in less than three or four seconds. It doesn't seem to draw an awful lot of power when it's asleep either; I never power the thing down, yet it still has plenty of battery life if I leave it in a cupboard over the weekend.
So, using MRAM to speed up booting is perhaps fixing a problem which shouldn't exist.
Actually this is a feature in ACPI S3 specification that is already supported by the beast server 2003 and should be supported in the upcoming 2.6 kernel.
As I mentioned in another post - if standby/hibernate was *reliable* (as in works 99 out of 100 tries) - I think it's possible that more people would use it.
However, my personal experience is that at *least* 1 of 10 times, the system will hang or otherwise fail to come back from standby/hibernate.
At which point I have to waste time waiting for reboot and I may as well just have left the system on with the monitor turned off.
Wolde you bothe eate your cake, and have your cake?
it stands to reason that once we have MRAM devices in practical use, the 'reset' button on today's boxes (read: instacrash/restart) will become a degauss memory/restart) button, thereby triggering a reboot of the system from a different set of MRAM sticks (assuming one stick holds kernel, drivers, software, docs ala Hard Drive while another is used for temp storage (read: modern RAM), so that if it necessary to reboot (i.e. cold reboot, it's gonna happen) it can be done without fragging the o/s and stuff, after all one doesn't reload the HD after resetting a box now do they, though some *nix gurus might
Logistical Chaos Officer http://www.slagg.org - LAN Gaming in Sarasota FL,USA