Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Dell Exec Calls HP's New 'Machine' Architecture 'Laughable'

timothy posted about 4 months ago | from the color-everyone-surprised dept.

HP 173

jfruh (300774) writes HP's revelation that it's working on a radical new computing architecture that it's dubbed "The Machine" was met with excitement among tech observers this week, but one of HP's biggest competitors remains extremely unimpressed. John Swanson, the head of Dell's software business, said that "The notion that you can reach some magical state by rearchitecting an OS is laughable on the face of it." And Jai Memnon, Dell's research head, said that phase-change memory is the memory type in the pipeline mostly like to change the computing scene soon, not the memristors that HP is working on.

Sorry! There are no comments related to the filter you selected.

Biggest problem (-1)

Anonymous Coward | about 4 months ago | (#47235477)

A reboot won't make it all good.

With persistent memory, the machine state gets messed up, you are so screwed.

Re:Biggest problem (3, Funny)

sjwt (161428) | about 4 months ago | (#47235501)

Just like all those CMOS chips that once you fuck up a setting their is no way at all clear them..

You can remove the CMOS battery for a while or (3, Informative)

mimino (1440145) | about 4 months ago | (#47235561)

You can remove the CMOS battery or move the Clear CMOS jumper or power on the PC with a special key pressed (depending on the motherboard manufacturer it can be CTRL, or ALT or something else, always well documented).

Old news, circa 2011 (0)

Anonymous Coward | about 4 months ago | (#47235819)

http://www.theregister.co.uk/2011/10/10/memristor_in_18_months/

... HP will target multi-layer memristor technology at the DRAM market in 2014/2015, with the SRAM market in its sights after that ...

... also at http://www.eetimes.com/document.asp?doc_id=1260374

... When challenged over the cost of the technology, which would be the barrier to competing against the high-volume flash memory market, Williams said: "On a price per bit basis we could be an order of magnitude lower cost once you get the NRE [non-recurring expense] out of the way" ..

Re:Old news, circa 2011 (3, Interesting)

Anne Thwacks (531696) | about 4 months ago | (#47235929)

once you get the NRE [non-recurring expense] out of the way

The entire cost of electronics is the NRE: look at your $800 iPhone - raw materials inside:

Three spoonfulls of oil to make the plastic bits.

Two spoonfulls of sand to make the silicon bits (includes the glass screen and fibreglass PCBs).

Not quite enough copper to make 2 inches of water pipe,

Not quite enough steel to make a table knife or fork.

Not much at all of quite a few other things

Way more than 2,000,000 man-hours of highly paid engineers' design time (if you include time to design every single component, including bought-in CPU, graphics, etc- remember to descend recurssively into the design of every single bit of logic, power disttribution, analog bits). Of course most has been amortized over the past 50 years, Apple only pays for the top layer.

If you start again from scratch, you might not need to go back to George Boole, or Aristotle, but you risk having to redevelop one hell of a lot.

Perhaps you shold meet a few engineers and talk to them.

Re:Old news, circa 2011 (2)

ColdWetDog (752185) | about 4 months ago | (#47236125)

Not those Non Reoccurring Expenses. He's talking about the cocaine and associated business costs with marketing and sales meetings.

Re:Old news, circa 2011 (2)

careysub (976506) | about 4 months ago | (#47236447)

Another way to look at it: the $800 iPhone 5S 64GB contains $210 of parts and cost $8 to assemble [cnet.com] , with giving an almost 300% mark-up. Laptop margins are usually 10% or less, Apple's laptop mark-ups are greater, around 30%. 300% is really remarkable.

Way more than 2,000,000 man-hours of highly paid engineers' design time (if you include time to design every single component, including bought-in CPU, graphics, etc- remember to descend recurssively into the design of every single bit of logic, power disttribution, analog bits). Of course most has been amortized over the past 50 years, Apple only pays for the top layer.

...

I guess we should count all of the hours spent in metallurgic and mechanical development since the dawn of the Industrial Revolution when considering the cost of car then?

Re:Biggest problem (2)

ledow (319597) | about 4 months ago | (#47235507)

Maybe we'll see a return to proper programming to go with this new technology, then. I doubt it, but maybe.

Re:Biggest problem (0, Funny)

Anonymous Coward | about 4 months ago | (#47235545)

It must be nice to be so perfect and odor-free.

Re:Biggest problem (2)

aliquis (678370) | about 4 months ago | (#47235655)

Political programming?

I will set var A to 5.

If it's:
*
*
*
You want I can do it.

Var A is 4.5 we'll try to make it 5 the next period.

Re:Biggest problem (4, Insightful)

K. S. Kyosuke (729550) | about 4 months ago | (#47235613)

"With persistent memory, the machine state gets messed up, you are so screwed."

Uh, have you looked into your computer recently? I believe you'll find either this little device called "an HDD" or this other little device called "an SSD". And people with those seldom get screwed.

Re:Biggest problem (0)

Anonymous Coward | about 4 months ago | (#47235635)

Yeah, and how many times have you just had to reformat the disk, because, really, it just won't recover ?

More than once, multiply that by millions of machines ?

Re:Biggest problem (2)

dreamchaser (49529) | about 4 months ago | (#47235679)

A handful of times maybe in over 20 years, and I keep backups that I rarely need.

Re:Biggest problem (0)

Anonymous Coward | about 4 months ago | (#47235993)

*rolls eyes*

your machine state is in your CPU and RAM
your persistent data is in your HDDs and SSDs

HP are trying to build hardware and an OS on top of it where this distinction no longer exists, with nonvolatile storage at "the speed of RAM", why not just store everything, both machine state and persistent data, in the one magical type of memory?

one of the many problems with this reasoning is exactly what you've failed to understand: with persistent memory, the machine state, as opposed to your donkey porn, gets messed up, and so you get to enjoy the brokenness

HP will doubtless counter with yet more pie in the sky "well, we could make a new breed of OS that is resilient in the face of corruption of critical structures', etc. for the last 20 years, HP labs have been layering bullshit on top of bullshit.

Re:Biggest problem (0)

Anonymous Coward | about 4 months ago | (#47236219)

one of the many problems with this reasoning is exactly what you've failed to understand: with persistent memory, the machine state, as opposed to your donkey porn, gets messed up, and so you get to enjoy the brokenness

*rolls eyes*

memset (0, 0, MAX_SIZE);

There, problem solved. Why do I keep reading Slashdot comments for technical topics?

RTFA (4, Informative)

tomxor (2379126) | about 4 months ago | (#47236337)

"With persistent memory, the machine state gets messed up, you are so screwed."

Uh, have you looked into your computer recently? I believe you'll find either this little device called "an HDD" or this other little device called "an SSD". And people with those seldom get screwed.

If you read the article [businessweek.com] from the previous slashdot story [slashdot.org] about HP's "The Machine", you will find that they are not simply trying to use memsistors to replace main memory, but that they are also trying to consolidate the storage memory and working memory into a single piece of memory, this is why it is considered to be substantially different memory architecture which also requires the OS to work a little differently too... if you are old enough think "Ram Disk"

The difference being that usually any stored data to be used by the processor has to first be loaded into working memory from the large slow storage memory... as i'm sure you are aware, which is why SSDs are so popular... but even NAND is many times slower than SDRAM, so the separation remains.

The idea is that if a sufficiently fast, dense, persistent and cheap type of memory can be found then the best of both can be consolidated into one. The concern of the OP is that issues affecting running state could affect the traditionally less dynamic stored state... Working memory is usually treated as volatile and disposable, and your block device is not, but the line is now blurred.

I think it's a reasonable concern, but one that is likely to be addressed by the OS, a less physical separation between what is running state and what is not would need to be implemented, but at the same time the advantages of not "loading" data need to be retained... making everything that goes into the running state duplicate would bring back the "loading" problem slightly.

Re:Biggest problem (0)

Anonymous Coward | about 4 months ago | (#47235999)

A reboot won't make it all good.

With persistent memory, the machine state gets messed up, you are so screwed.

Not such a big deal, really... as long as you engineer the machine so that it still bootstraps off CMOS like right now. The process would essentially be the same.

OS Lock In (0)

Anonymous Coward | about 4 months ago | (#47235479)

MS must be loving this news!

Now we'll need a new form of ReactOS and/or WINE to reverse engineer this new OS.

Good times.

Re:OS Lock In (3, Insightful)

Quarters (18322) | about 4 months ago | (#47235765)

Do you truly, honestly, I mean...REALLY believe that Microsoft expends any time at all even thinking about ReactOS or WINE, let alone worrying about the .00000000000001 of a fraction of a portion of a negligible amount of a percent effect it might, MIGHT have on their bottom line?

Seriously, answer seriously, please.

Re:OS Lock In (0)

Anonymous Coward | about 4 months ago | (#47236165)

M$!

Re:OS Lock In (2)

fizzer06 (1500649) | about 4 months ago | (#47236533)

This might be the year of ReactOS on the desktop.

Prist Fost! (-1)

Anonymous Coward | about 4 months ago | (#47235481)

w000tttt!!!! elelven!!!!!!

HP should liquidate (2)

BorgDrone (64343) | about 4 months ago | (#47235483)

Yeah, maybe HP should shut down and give the money back to the shareholders. Right ?

Re:HP should liquidate (1)

aliquis (678370) | about 4 months ago | (#47235661)

Did the buy out of Dell happen?

Because if so they kinda followed through with that idea themselves at least.

Not invented here syndrome (0)

Anonymous Coward | about 4 months ago | (#47235495)

It is kinda similar to not in my backyard syndrome.

Re:Not invented here syndrome (0)

Anonymous Coward | about 4 months ago | (#47235529)

Yeah... At first they laugh at you...

All is good and well in computing. (0)

Anonymous Coward | about 4 months ago | (#47235499)

Quite frankly I am happy this is happening, more competition to create the Next Big Thing is glorious.

There are so many great new tech that is around the corner, and the more competitors the better.

I'll just be happy that x86 dies the death it should have died many years ago.
It isn't even a horse any more, they are beating the bloody, dirty pulp that remains.

Magical States? (0)

Anonymous Coward | about 4 months ago | (#47235509)

Probably not a good idea to be dissing magic around computer wizards.

Mr Dell's just upset (5, Insightful)

The123king (2395060) | about 4 months ago | (#47235511)

No-one's interested in his shitty computers anymore

Re:Mr Dell's just upset (4, Informative)

Lumpy (12016) | about 4 months ago | (#47235943)

Nor HP.. HP's quality has tanked hard as well. Most of their Mexico Assembled crap fails quickly. 5 desktops quad i7 top of the line HP boxes, 3 of them had problems that required a major repair like mother board replacement.

It seems that all the computer makers are just building low grade dog food these days.

Re:Mr Dell's just upset (0)

Anonymous Coward | about 4 months ago | (#47236119)

HP is orders of magnitude worse.

This is how I know HP is on to something (3, Insightful)

Anonymous Coward | about 4 months ago | (#47235517)

If Dell has to misrepresent what HP is doing in one breath while disproving that misrepresentation in the next, just to have a straw man to poke fun at, then Dell must be a little scared.

Re:This is how I know HP is on to something (1)

Anonymous Coward | about 4 months ago | (#47236021)

If Dell has to misrepresent what HP is doing in one breath while disproving that misrepresentation in the next, just to have a straw man to poke fun at, then Dell must be a little scared.

Bingo. Personally I'm not on either "side", I believe the first guy who actually brings something to market which can be performance tested. But I find little credibility in his claims, or any time Company X bashes on Company Y's not-yet-invented machine by pointing to their own (and less likely) not-yet-invented machine.

Call it: THOLEAN TECH (0)

Anonymous Coward | about 4 months ago | (#47235541)

Because of the phase change. You see. Tholean. Right. Thol. E. En. Is this thing on?

The definition of innovation (1)

Anonymous Coward | about 4 months ago | (#47235565)

HP, despite leadership's best efforts throughout the years, still does legitimate innovation. Dell has never done the whole innovation thing.

Re:The definition of innovation (0)

Anonymous Coward | about 4 months ago | (#47235581)

Hey, you take that back! Dell were the market changing innovators of green plastic components!

I must respectufully disagree, (1)

josquin9 (458669) | about 4 months ago | (#47236551)

. . . or at least suggest that you qualify your statement.

At it's inception, Dell was quite innovative. But that innovation was limited to business practices, not engineering or technology. Compaq got the ball rolling, but Dell developed the production and marketing models that brought the price of usable desktop computers down to the sub-$1000 level. This was probably as instrumental in putting a computer on every desktop as anything Bill Gates did. Other manufacturers copied and improved upon the model later, but Dell's decisions not only made computers more affordable, but also introduced to a generation concepts that are now considered mainstream (CPU, RAM, etc.), but which had been considered indecipherable techy arcana. I believe this significantly increased computer adoption, simply by demystifying the strange beige boxes.

These may not have been technological innovations, but they were definitely innovations, and led to the kind of "creative destruction" so often given lip service by conservatives (whose actual practices are mostly about maintaining market stability and the fortunes of those who have already won them, rather than innovation.)

Dell can have no valid opinion on this. (4, Insightful)

Required Snark (1702878) | about 4 months ago | (#47235571)

Dell is a reseller. They do not invest in any of the fundamental technologies like CPUs or Operating Systems. They have no design expertise in virtual machines like the JVM. They don't do chip design or fab. They have never been in any of these businesses.

HP has a long history of OS and CPU design, including their own computers with a proprietary architecture. Not all of their designs were successful, since they were co-designers of the Itanium with Intel. So HP has the exactly opposite corporate background the Dell.

Why would anyone pay attention to what a Dell talking head has to say?

Re:Dell can have no valid opinion on this. (0)

Anonymous Coward | about 4 months ago | (#47235687)

Because business by Dell by the thousands, and just want generic x86 boxen. Execs know the brand, so when someone from Dell speaks, your superiors will listen to them more than you. That's why.

Re:Dell can have no valid opinion on this. (2)

grep -v '.*' * (780312) | about 4 months ago | (#47235695)

Why would anyone pay attention to what a Dell talking head has to say?

DUUUude, that's harsh.

Re:Dell can have no valid opinion on this. (1)

Jim Sadler (3430529) | about 4 months ago | (#47235707)

It is obvious that HP has been important in the development of computers. Dell seems to be concentrated on sales and producing computers that tend to be clunky and in my opinion crowded enough internally to be durable and unable to use third party hardware. I simply do not enjoy Dell's desktops.

Re:Dell can have no valid opinion on this. (1)

TapeCutter (624760) | about 4 months ago | (#47235717)

Putin doesn't need Dell's "help" either, just watch his eyebrows [youtube.com] hit the ceiling when the Dell CEO offers it. Eyebrow event occurs ~1:00.

Re:Dell can have no valid opinion on this. (0, Flamebait)

dfghjk (711126) | about 4 months ago | (#47235763)

"Dell is a reseller. They do not invest in any of the fundamental technologies like CPUs or Operating Systems. They have no design expertise in virtual machines like the JVM. They don't do chip design or fab. They have never been in any of these businesses."

They, in fact, have been in all those businesses. Dell did their own Unix and they supported both gcc and X11 development. They invested in CPU and chipset technology; several ex-Dell employees left to found Centaur. They work closely with all these technologies. They have to.

HP has the exact same corporate background as Dell; their old histories are irrelevant. HP was an instrumentation company once upon a time while Michael Dell was an Apple II hacker. So what?

Ignorant people should learn to keep their mouths shut.

Re:Dell can have no valid opinion on this. (0)

Anonymous Coward | about 4 months ago | (#47235957)

They, in fact, have been in all those businesses. Dell did their own Unix and they supported both gcc and X11 development.

Oh wow yeah, Dell UNIX. A UNIX so obscure that even Wikipedia doesn't care about it. I've seen more people using A/UX than Dell UNIX.

They invested in CPU and chipset technology; several ex-Dell employees left to found Centaur.

Woo? HP have actually designed and built their own CPU's, and still to do this day design and build all sorts of ASICs for things like RAID controllers.

The fundamental difference is that HP actually make things, and Dell just tinker around the edges without ever getting their feet wet. Oracle are more innovative than Dell.

Re:Dell can have no valid opinion on this. (2)

ColdWetDog (752185) | about 4 months ago | (#47236157)

Oracle are more innovative than Dell.

Now you're just being mean.

Re:Dell can have no valid opinion on this. (0)

Anonymous Coward | about 4 months ago | (#47235963)

> They, in fact, have been in all those businesses. Dell did their own Unix and they supported both gcc and X11 development.

Dude, that was 20 years ago. I remember when their version of SVR4 was arguably the best one out there, but it was SVR4 for chrissakes!
They gutted that whole department long, long ago which is why those guys left to start Centaur in 1995.

Re:Dell can have no valid opinion on this. (4, Interesting)

pepty (1976012) | about 4 months ago | (#47235893)

HP has a long history of OS, CPU, and other types of tech design, but they lost a lot of that when they spun off Agilent. Since then HP's budget for research, not to mention the researchers/departments themselves, have been slashed. They are not down to Dell levels of R&D yet, but that seems to be the trend.

Re:Dell can have no valid opinion on this. (1)

timeOday (582209) | about 4 months ago | (#47236571)

Why talk in generalities about brands? Let's just compare the work. What project is Dell working on that's comparable to "The Machine"? Nothing.

Re:Dell can have no valid opinion on this. (2, Interesting)

Lumpy (12016) | about 4 months ago | (#47235949)

"HP has a long history of OS and CPU design, including their own computers with a proprietary architecture."

and nobody works there anymore that does that, They fired all the high paid specialists years ago.

The HP of today is not even worthy to stand in the shadow of the HP of yesterday.

Re:Dell can have no valid opinion on this. (0)

Anonymous Coward | about 4 months ago | (#47236017)

this is a tangent, but:

HP has a long history of OS and CPU design, including their own computers with a proprietary architecture. Not all of their designs were successful, since they were co-designers of the Itanium with Intel

Actually, the extent of HP's involvement in Itanium was providing the (excellent, IMHO) PA-RISC architecture as a baseline. Pretty much everything shit in Itanium was thanks to Intel.

Here's what I don't care about: (2, Insightful)

Anonymous Coward | about 4 months ago | (#47235575)

What an executive from Dell, a company that is almost single-handedly stifling innovation in the computer industry by continuing to push enormous volumes of generic wintel garbage out onto the market to the exclusion of anything else/new/better/etc, has to say about innovation.

Dell desperately seeks relevance in today's market (1)

Anonymous Coward | about 4 months ago | (#47235579)

Dell is rendered irrelevant nowadays. So they are looking for publicity to stay in the minds of the few who still like to hold on to the crap of yesteryears like my attachment to toshiba libretto mini laptops.

Bad mouthing others is often a good way to get publicity. They will be rendered mute by industry in a few weeks like qcomm's 64bit outcry - necessarily pointing out -"waa waa, he did it while I couldn't".

When did dell get any innovative stuff out ? Their business model in the beginning was probably the only true innovation. After that cheapness coefficient is the only discerning factor in their persona.

Dell argument is wrong (2)

brysiek (468731) | about 4 months ago | (#47235583)

It is not CPU and Memory being the two main core components of modern computing fabric. Instead, it is the inter-connect and memory, and with these two, new high performance operating system would have to be developed.
If you look at today's data center processing vast amount of data, you can see that most of the space is not taken by servers with CPUs.

Re:Dell argument is wrong (1)

dbIII (701233) | about 4 months ago | (#47235645)

If you look at today's data center

Most of the space is typically taken up with disks but that's irrelevant, just like it was in the 1960s, the important thing is what you actually do with the data and not how much you can hoard.

Instead, it is the inter-connect and memory

While for some tasks I would really like to see direct connections between CPUs in different cases/racks and vast amounts of shared memory so a cluster can be treated as a single machine in more than an abstract sense I think I'm in the minority (look at all those VMs) and the latency tradeoffs are going to suck for other applications. If the physical distance to that other CPU is metres and you've got to handle switching to a lot of CPUs there is a price - as already seen on a small scale where the multi-way AMD and Intel CPUs are a full 1GHz or more slower than their single socket versions.
So in my opinion, while such a focus would be something I'd really like to see I don't think the mainstream is going to go for it due to the tradeoffs.

Hopefully you meant CPU to CPU interconnection such as "hypertransport" and not mere node to node networking.

Re:Dell argument is wrong (0)

Anonymous Coward | about 4 months ago | (#47236061)

Most of the space is typically taken up with disks but that's irrelevant,

That's actually directly relevant to this tech discussion.

The point is that you can have memory which works at the speed of RAM but has capacity on par with SSD's and HDD's. It's a way to eliminate a lot of the bottlenecks involved with shuffling data between slow, long-term storage units into the fast, but small, short-term 'workspace' units.

Put another way, it's like having a RAMdisk large enough to put your entire HDD on. And that has many, many applications and benefits.. people just don't do it much because RAM is so costly and limited in space.

Re:Dell argument is wrong (0)

drinkypoo (153816) | about 4 months ago | (#47235731)

It is not CPU and Memory being the two main core components of modern computing fabric. Instead, it is the inter-connect and memory,

How far will you get without your CPU? You can't execute one instruction. You can't even POST. The CPU and memory are the two main core components of modern computing fabric.

What is the Dell CEO supposed to say? (5, Interesting)

Mysticalfruit (533341) | about 4 months ago | (#47235603)

HP is their competitor. HP just announced that they're working on something that even if the entire thing doesn't come to fruition, likely some part will and it will change the computing landscape. Understand, this announcement is pointed directly at Dell's share holders.

Best case scenario HP actually pulls it off and they've got some radically fast system running something that looks like Linux.
Mid case scenario, they figure out how to make memsistors at scale and then sell licences for everybody to make blisteringly fast SSD's, etc. Then others come along and figure out how to put the pieces together. HP makes out like a bandit in royalties, etc.
Worse case, nothing comes out of this. HP shrugs, files a whole pile of patent applications. Someone else takes bits and pieces of it (like IBM) and does cool things with it. In all three cases HP is going to be enhance their IP portfolio and possibly make their stock worth more.

All of those scenarios are bad for Dell. Dell doesn't do fundamental science. They design motherboards that use components supplied by everybody else and crank out cheap computers. If scenario #1 comes true... HP is NOT going to sell any of this to Dell, cutting them out of the market. If scenario #2 comes true, HP is going to get these components at a price that Dell can't compete with. If the last scenario comes true, Dell still ends up being a VAR like everybody else and HP racks in royalties.

The CEO of Dell is almost obligated to thrown cold water all over this, otherwise Dell shareholders are naturally going to ask if this announcement is going to make Dells stock worth less and/or worthless.

Re:What is the Dell CEO supposed to say? (0)

K. S. Kyosuke (729550) | about 4 months ago | (#47235643)

some radically fast system running something that looks like Linux.

If it looked like Linux, it most certainly wouldn't be "radically fast". Or at least "radically faster than Linux on commodity HW". You really have to throw away all your preconceptions - and your existing SW - if you want to really reap the benefits of huge non-volatile random access storage. Why would you design screaming-fast hardware and then cripple it with inadequate software? Especially if it's new and expensive technology.

Re:What is the Dell CEO supposed to say? (1)

wjcofkc (964165) | about 4 months ago | (#47235699)

Just because something looks like Linux or any *nix does not mean that it can't actually be fundamentally and radically different under the hood. The likes of Bash, csh, zsh, etc. are still the most powerful ways to interact with a computer - by far and away. The Unix console has survived as the ultimate interface since the sixties for a very good reason. Extending a shell to take advantage of new hardware functionality would actually make sense in this case: it's powerful, and admins would already know how to use it. Also, it doesn't sound like HP is exactly going straight for the desktop with this. Unless you want a refrigerator sized machine in your house. If you are envisioning running KDE on this thing, you are thinking about it wrong.

Re:What is the Dell CEO supposed to say? (1)

K. S. Kyosuke (729550) | about 4 months ago | (#47235767)

Just because something looks like Linux or any *nix does not mean that it can't actually be fundamentally and radically different under the hood.

Linux is a un*xy system. It's all built around the notion of byte streams, also known as files, organized in a hierarchical fashion, optimized for streamed (or at least semi-random, block-sized) access. This new hardware brings the promise of persistent heaps. How exactly do you propose to design an OS for that, keeping the benefits of persistent data objects, while running applications working on serialized data on top of that? That "fundamentally and radically different" thingy underneath the file-emulating API would get horribly screwed, performance-wise.

Re:What is the Dell CEO supposed to say? (4, Insightful)

wjcofkc (964165) | about 4 months ago | (#47235833)

If you read the original article about the technology, they have competing OS development teams. One of them is working on a new Open Source "Machine OS", another team is working on developing a modified version of Linux to take advantage of what the platform could potentially offer. As long as they are bothering to do that at all, I would say they know what they are doing and have a working answer to your question:

How exactly do you propose to design an OS for that, keeping the benefits of persistent data objects, while running applications working on serialized data on top of that?

Re:What is the Dell CEO supposed to say? (1)

K. S. Kyosuke (729550) | about 4 months ago | (#47235873)

I've read it. But the Linux endeavor is really nothing else than detox. There is no "working answer" for that, it will be like running legacy DOS software in DOSBox on a Core i7 machine. Just because people are doing just that doesn't mean that Intel is advocating writing new DOS software as the preferred way of software development for Core i7.

Re: What is the Dell CEO supposed to say? (2)

Trinn (523103) | about 4 months ago | (#47235839)

Its called mmio, mmap() specifically. Linux already has xip support on some platforms as well. This is all under the hood too, the libc could be redesigned, or insert your favorite language here. I agree that writing code optimized for it might be a bit different but its not that different than writing for an all-sram platform like say the old palm.

Re: What is the Dell CEO supposed to say? (1)

K. S. Kyosuke (729550) | about 4 months ago | (#47235883)

I'm perfectly aware of mmap. I'm just not sure if anyone has ever run large scale automated memory management on something like that (mmap over TB-sized files, for example). The solutions we have for that, like Azul's C4, have some pretty strict requirements as to what the HW has to be able to do.

Re:What is the Dell CEO supposed to say? (0)

Anonymous Coward | about 4 months ago | (#47235701)

If they can write a Unix like OS(i.e. "something that looks like Linux") that can take advantage of this system then it will give them an advantage, since people, especially high performance computer users already know how to use Unix APIs. The more familiar to traditional high end servers and supercomputer users it is the faster they will generate customer interest. Given this they are likely to borrow more heavily than just the Unix standards if they can, if they are similar enough to Linux to use Linux desktops then they don't need to write their own and the same applies to any other software that they might need. Since the hardware is theirs they do not need to own the software to keep the market theirs, so why not ?

Re:What is the Dell CEO supposed to say? (1)

K. S. Kyosuke (729550) | about 4 months ago | (#47235781)

What if ditching Unix would give you up to two extra orders of magnitude in performance? Would you still clamor for a Unix-compatible API?

Re:What is the Dell CEO supposed to say? (0)

Anonymous Coward | about 4 months ago | (#47236281)

Why write everything from scratch right up from the command line text editors to the convenient graphical user interfaces, when you do not have too? Clamour is not the right word, it would be interesting to see a new scratch built OS architecture, but I think they will include enough compatibility to avoid having to recreate everything from scratch. The underneath may be different on the inside and the high performance programs may be completely different too, but most of the programs you are using day to day don't need the performance anyway so why not limit the costs?

In many cases "can buy and get two times the performance now" will be bought but "can buy then re-write all your software from scratch to get 200 times more performance" wont. Certainly there is a market for the latter but, where depends on the need for performance and the cost of rewriting the program, as well the difference between the cost of the new machine and it's rivals when buying replacements for current kit. If it costs the same and gives even slightly better performance per watt without requiring re-writes then it will sell like hotcakes and everybody will want one, "could be better" is irrelevant. If it costs a tone more than it's rivals and requires a complete rewrite of software just to get it to run, as well as needing the technical staff to relearn how to use it from scratch, then only those needing the performance and for whom the re-write is not an excessive cost or delay will bother to buy it, deep pockets but low in number. Of course the opportunity to bypass any compatibility layers, and do a rewrite to gain such a benefit, after you have the machine and already gained some advantage, would add significantly to the value, but programmers are not cheep.

Re:What is the Dell CEO supposed to say? (0)

Anonymous Coward | about 4 months ago | (#47236001)

since people, especially high performance computer users already know how to use Unix APIs.

And MPI too. When you program to MPI, you don't care what the fuck the underlying OS looks like, you can recompile and be running on CNK, recompile again and be running on L4, recompile again and be running on Linux, it doesn't fucking matter, because the only API you ever see is MPI.

When you kick filesystems to the kerb, the *only* thing Unix gives you is a horrible sockets API. MPI is a much nicer API for message passing, and a much nicer API for file access, so I really wouldn't miss anything from unix.

Dell is a privately held company. (2)

Reibisch (1261448) | about 4 months ago | (#47235743)

Announcements from executive leadership to ownership are made via boardroom table, not to reporters.

If you want to make an argument that Dell's 'announcement' was made to Dell customers or partners, you might be able to make a case. But the thought that they're 'announcing' this to rally support of shareholders is laughable.

Re:Dell is a privately held company. (0)

Anonymous Coward | about 4 months ago | (#47236173)

Announcements from executive leadership to ownership are made via boardroom table, not to reporters.

If you want to make an argument that Dell's 'announcement' was made to Dell customers or partners, you might be able to make a case. But the thought that they're 'announcing' this to rally support of shareholders is laughable.

There are a lot of shareholders who are neither owners nor board members. They are publicly traded, and perception by the public sector can and does influence stock prices.
And I think you don't quite understand- Dell isn't announcing it to "rally" shareholders, but to try and put doubt in the minds of those who might buy HP stock, and to calm the fears in those who hold or are thinking of buying Dell stock.

The Stock Market is in many cases just as much of a matter of PR as it is of company performance.

Re:What is the Dell CEO supposed to say? (2)

dfghjk (711126) | about 4 months ago | (#47235799)

- Swanson is not Dell's CEO
- Dell is under no obligation to comment, they could opt to do what all of HP's other competitors do
- Dell is privately held, it has no shareholders
- modern HP has proven incapable of delivering tech that leads
- other companies already ARE selling NV Ram technologies into storage markets, HP won't be getting royalties on this
- HP doesn't have to sell to Dell, but they have to sell to somebody. They won't establish anything as standard on their own
- Building a business on crappy products and loads of IP worked great for TI, they're just like Dell

Re:What is the Dell CEO supposed to say? (1)

ultranerdz (1718606) | about 4 months ago | (#47235959)

TI?

Re:What is the Dell CEO supposed to say? (2)

wisnoskij (1206448) | about 4 months ago | (#47235849)

"Worse case, nothing comes out of this. HP shrugs, files a whole pile of patent applications. Someone else takes bits and pieces of it (like IBM) and does cool things with it. In all three cases HP is going to be enhance their IP portfolio and possibly make their stock worth more."

Aren' patents great. Even if you fail to invent anything that works you can just file a general patent for the technology and claim royalties on a design that someone else actually gets to work, in perpetuity.

Re:What is the Dell CEO supposed to say? (1)

Askjeffro (787652) | about 4 months ago | (#47236489)

1. Dell is now a private company, so no need to worry about public shareholders anymore. 2. Swanson is not Dell's CEO, he is head of software.

Simple explanation: John Swanson is scared. (4, Interesting)

K. S. Kyosuke (729550) | about 4 months ago | (#47235631)

Or an idiot. Or a scared idiot.

The notion that you can reach some magical state by rearchitecting an OS is laughable on the face of it

Why, thank you, Captain Obvious! It's not about rearchitecting an OS, it's about matching SW to the HW. For ages, we've had the distinction between block-addressed devices with streamed access and byte-addressed devices (mostly DRAMs) for low-latency. Virtually all our software is impedance-matched to that idea! I believe the only thing remotely close to how a machine with huge persistent RAM should (would?) work are those nice Azul boxes, with zero-pause automated memory management even on 500GB+ heaps. Those machines still use RAM and have disk I/O for ordinary data manipulation, but I'm convinced that had the Azul people had non-volatile RAMs at that time, they would have gone for persistent objects. It's such an obvious idea! No more serializing and deserializing for disk I/O (except for backups, of course), performance on the order of millions of transactions per second. Obviously the price is that you absolutely have to rewrite the software bottom-up, otherwise all that extra performance potential gets lost.

Re:Simple explanation: John Swanson is scared. (2)

Lumpy (12016) | about 4 months ago | (#47235965)

"Obviously the price is that you absolutely have to rewrite the software bottom-up, otherwise all that extra performance potential gets lost."

Which is a GOOD THING (tm). The current state of software quality is horrid, anything to force a rewrite will be a very good thing.

Re:Simple explanation: John Swanson is scared. (1)

CBravo (35450) | about 4 months ago | (#47236713)

They should not only redesign software for CPU architecture. Most software is created for a single cpu and a single memory space. In real life we have multiple processors, multiple kinds of memory (cache, ram, disk/ssd, raid, san, distributed file systems), network interfaces between server and client (what do you consider 'an application' on the internet?).

And while we are at it: We have issues with software reuse, bugs (in general) and testability, security. Software development is in the pre-industrial age, afaiac.

Re:Simple explanation: John Swanson is scared. (1)

Anne Thwacks (531696) | about 4 months ago | (#47235969)

What is needed to process large amounts of data is a massively parallel data-flow architecture - something resembling a hardware implementation of SQL. The ICL DAP is an early example. The Cell processors in your Playstation 3 are a half baked attempt at the same thing. You would probably still want a conventional processor to supervise it, and probably to compile the programs.

It is not difficult to make one of these using conventional; silicon.

It is hard as hell to get funding and sell it. If you actually want to make one, drop me an email (and several billion dollars). Yes, I have worked on this stuff before.

Re:Simple explanation: John Swanson is scared. (1)

K. S. Kyosuke (729550) | about 4 months ago | (#47236109)

How well does dataflow architecture with random access data structures? Indices etc.? For that matter, how would it run something like AllegroGraph? (Given that this is another interesting area of application for such machines.)

Memsistor are cool (1)

AndyKron (937105) | about 4 months ago | (#47235641)

I think memsistors will give us human-like computers

Re:Memsistor are cool (1)

wonkey_monkey (2592601) | about 4 months ago | (#47235837)

Because...?

Re:Memsistor are cool (1)

Artifakt (700173) | about 4 months ago | (#47235989)

Cause now everyone knows what Positrons are, Asimov's positronic brain sounds dated.

Phase Change is the same but faster (2)

Karganeth (1017580) | about 4 months ago | (#47235651)

Memristors are a fundamental change in computation. Fuck dell and their bullshit spewing CEOs. Burn in hell dell.

Re:Phase Change is the same but faster (1)

Anne Thwacks (531696) | about 4 months ago | (#47235995)

No they aren't. Memristors are a trivial change in how you implement a low-level feature. Its like whether you use polythene or polycarbonate for your capacitors.

While I would be quite happy for Dell to burn in hell, taking i86 architecture with them, a new computer architecture is a completely different plot from new implementations of memory or a new software design. Memristors are not even content addressible memory - which have been done in silicon, and shown to make text searching and jump tables (case statements) thousands of times faster, but no one will buy. (NIH? risk averse management? Decisions taken by non-tech people, PHBs)

History tells us that writing a new OS for new hardware is pretty much the best way to ensure your project fails. Do one of the other, not both.

Re:Phase Change is the same but faster (0)

Anonymous Coward | about 4 months ago | (#47236123)

Oh realy? Cheap sram is a trivial change? I don't thinkg you understand that you can very well ditch hard drives if they're as fast and reliable as promised.

I wonder if this applies (3, Insightful)

skovnymfe (1671822) | about 4 months ago | (#47235659)

I wonder if this applies: First they ignore you. Then they ridicule you. And then they attack you and want to burn you. And then they build monuments to you.

Re:I wonder if this applies (2)

wonkey_monkey (2592601) | about 4 months ago | (#47235863)

First they ignore you. Then they ridicule you.

And most of the time that's where it stops, because the idea was ridiculous.

uh no (4, Interesting)

drinkypoo (153816) | about 4 months ago | (#47235673)

"The notion that you can reach some magical state by rearchitecting an OS is laughable on the face of it," John Swainson, head of Dell's software business, told reporters in San Francisco Thursday when asked to comment on the work.

Well, sure, you also have to rearchitect the hardware, which is what HP is talking about. John Swainson is an idiot. Sadly, the richest idiots with the best-connected families fail upwards rather than downwards. This is why we can't have nice things.

Re:uh no (1)

fleebait (1432569) | about 4 months ago | (#47235937)

It might be just a little more than just a game changer.

Stop thinking about computers as boxes with wires, screens and disks, and start thinking about building the nervous system of a human being. Our bodies use distributed computing all over the place, with the vagus nervous system for the organs, with their own chemical memories, and feedback loops, the localized muscle memory systems for arms, legs, fingers, locally stored programs that run semi-autonomously.

If you read about memristors on Wikipedia, you can begin to see the possibilities of interfacing with biologic systems, and the newer bioligic chemical sensors within the organs, and appendages. Distribute local semi-dedicated processors with the distributed memory systems, and now we're talking about leaps ahead for automotons, and robotics. Who needs a stupid file oriented operating system, when the information needed for a process is stored locally.

Unix is so yesterday, as well as any other file orientated storage system.

How do you organize your brain? Do you have file cabinets, with tabs, disks? pictures? No, it's some sort of random access sensory system that relates to previously accessed information. Something like the memristors they are talking about.

It's coming down to defining the complete application, before building the actual machine itself.

I imagine early prototypes may be in a metal box with wires, but interface is going to be a new problem. Most likely all fibre connections before connecting directly to sensors and embedding sensory processing at the sensor itself -- -- and so on.

Re:uh no (1)

drinkypoo (153816) | about 4 months ago | (#47236079)

How do you organize your brain? Do you have file cabinets, with tabs, disks? pictures? No, it's some sort of random access sensory system that relates to previously accessed information. Something like the memristors they are talking about.

Memristors are predominantly a way to build things we already know how to build, but more efficiently. No one knows how the brain stores information. It is known that it is possible to create an index to information in your brain by imagining file cabinets, or rooms, or some other sort of containers. I forget things all the time, but my PC doesn't.

Dell's research head (0)

Anonymous Coward | about 4 months ago | (#47235723)

Those are the funniest three words I read today.

Re:Dell's research head (0)

Anonymous Coward | about 4 months ago | (#47235741)

Don't knock him, it takes a lot of research to find the latest and cheapest Chinese manufacturer who can shave 0.01c off of your margins. Particularly when they're down a back alley in Hangzhou and you don't have a map.

I for one... (3, Funny)

Alejux (2800513) | about 4 months ago | (#47235745)

...am super excited to see what kind of algorithms and applications could benefit from this kind of architecture: artificial intelligence, computer vision, ray-tracing, etc...

laughable? (0)

Anonymous Coward | about 4 months ago | (#47235785)

"The notion that you can reach some magical state by rearchitecting an OS is laughable on the face of it."

This is a very silly comment to make.What is laughable is the notion that today's software architecture somehow had reached some pinnacle of optimization.
Dell know how to assemble commodity-grade components spun off from the original IBM pc architecture. Why WOULD they know any architectural vision if it smacked them in the face...

The only laugable thing here (0)

Anonymous Coward | about 4 months ago | (#47235813)

is DELL making comments about HP.

Link to animated gif here (1)

the_saint1138 (1353335) | about 4 months ago | (#47235871)

Articles should include a link to the relevant video.

I found this gif of the event: http://stream1.gifsoup.com/vie... [gifsoup.com]

You're welcome.

Follow the money. (1)

ISoldat53 (977164) | about 4 months ago | (#47236141)

Of course the software guy at Dell is going to make noise about anything that competes with Microsoft. If it wasn't for them he wouldn't have a job.

BeOS did magical things (0)

Anonymous Coward | about 4 months ago | (#47236197)

Yes, magical things can be done at the OS level with some smart boffins - look at BeOS. I guess I am one of the view who who had a PC in the 1990s sporting Win95 that could barely play a two inch video clip without freezing. Meanwhile, on the same hardware BeOS could run clip after clip at the same time and never miss a beat! I wish I could remember how many videos and MP3s I had to load to bring it to its knees!

Having said this, as a former employee of the morally bankrupt HP it will be cold day in hell I give them another dollar.

Load More Comments
Slashdot Login

Need an Account?

Forgot your password?