Slashdot Mirror
Intel Plans to Overhaul Chip Architecture
Carl Bialik from the WSJ writes "Intel is planning to announce an entirely new chip architecture later this month at the company's developer forum, the Wall Street Journal reports. The company isn't discussing details yet, but it's expected that Paul Otellini will discuss a 'technology foundation designed from scratch to improve energy efficiency and make it easier to add more than two processors.'"
One thing the article didn't make clear is what exactly Intel means by "A New Chip Architecture". i.e. Do they mean a new architecture as in the Itanic (but low power!), or a new chip architecture as in, "We're ditching the 20 stage pipeline in exchange for a more reasonable 6 stage pipeline, swapping out most of the control circuts for those from our StrongARM line, and rewriting the microcode to execute all of the Pentium instructions on a simple, low power RISC core."
While they could go either way, I hope they've learned from the Itanium and EM64T debacles that they should stick with a compatible microcode. Leave the super-instruction sets to the MIPS and SPARCs of the world.
Javascript + Nintendo DSi = DSiCade
Part of the roadmap Jobs was talking about?
You remember back in the day when processors had only one core?
Slices, dices, eats your lunch.
Does this mean a new architecture like not being x86? Is this for Apple?
It'll also likely contain whatever TPM provisions that they've decided to spring on consumers, under the guise of having great new "features" and "advantages." They're always eager to talk about what you -can- do with their hardware, but relatively silent on what they're going to make sure you -can't- do.
Related to the Apple switch? Anyone?
Those who can, do.
Those who cannot, teach.
Those who think they can but cannot, manage.
Conroe according to Anandtech...4 92
http://anandtech.com/cpuchipsets/showdoc.aspx?i=2
HJ
Who wants to bet that the announcement includes a integrated memory controller? I wouldn't be suprised if they just licenced Opteron technology from AMD; it would be alot cheaper than developing their own. Although, they could always just outright steal it.
Bugs are just features that have been fixed.
Any thoughts on if this was prompted (in part) by apple, and their craze for multiprocceser computers?
Or is it just me experiencing this was announced oh-so-many times?
I mean- it might be what the market really needs - less power with bigger integration- but we've known that for at least um- 20 years?
--
http://www.linklike.de.vu/
Isn't this like, 15 years too late?
On NPR this morning, they mentioned that Intel had said that a typical PC user wouldn't notice any change as a result of this new architecture. So one presumes this means no major instruction set revisions or anything.
www.sjbaker.org
What is it about computers that makes people draw analogies to automobiles?
This is kinda funny in two ways..
'technology foundation designed from scratch to improve energy efficiency and make it easier to add more than two processors.'
Not overheard anywhere: "We are peeking through a knothole in AMD's fence and seeing what they are up to.
Nitpick: "The company isn't discussed details yet"
The proper word is ain't.
A feeling of having made the same mistake before: Deja Foobar
One has to wonder if Apple had any 'insight' to these plans when they signed the deal.
Perfecting Discordia
www.stevenvansickle.com
1. Redesign chip architecture 2. ???? 3. ???? 4. Profit!
Scott
considering remotely exploitable cache timing attacks against all software implementations of AES and other S-Box constructions will Intel grow some brain cells and put crypto directlty into the core?
i don't think they are that smart...
mac switches to intel :-p
intel switches to PPC ?
The war with islam is a war on the beast
The war on terror is a war for peace
Meanwhile, Intel's desktop dual core chips seem to offer much more aggressive pricing at this time. AMD's lowest price dual core chip, the X2 4200 is almost twice as expensive as Intel's lowest cost dual core processor. However, an interview with three AMD execs on PCPerspective.com claims that "AMD would eventually have lower priced Athlon X2 processors via the waterfall effect in the future".
As we all know, the Pentium 4 is a pretty goofy, shlocky design. The Pentium M is good, but it's essentially a Pentium Pro. That's 10 years old.
Intle NEEDS to prove that they can still make a good x86 chip from "scratch".
Yeah, I also remember when they had a 4.77 MHz processeor and no pipeline.
And guess what? Those days SUCKED.
Intel announced some fud about EPIC, and except for fujitsu who kept Sparc alive despite Sun's layoffs this FUD wiped out the entire market.
Methinks they saw the power of this approach and if the last round killed 4 leading nplayers, this round will kill off the remaining 2 (IBM & AMD).
I hope their new logo isn't as easily confused with a feminine hygine product.
Performance per watt? Notice now Intel is singing the exact same tune that Apple is? I'm not saying that it's being made specifically for Apple, but clearly Steve Jobs looked at the roadmap and, since Intel wants something new, saw a common goal that he could pursue.
This really makes me wonder if these new chips be the foundation of the Intel based Macs. Considering the talk about power consumption by Steve Jobs during the announcement, it seems pretty likely. Macs don't need all the old 16 bit support Windows relies on to boot and possibly run.
"Bishops and Bookies live off the irrational hopes of mankind." Bertrand Russell
Sun's Niagara is coming...
= Sun_Micro_Set_To_Release_Eight_Core_Niagara_System s&story_id=37716
http://www.newsfactor.com/story.xhtml?story_title
I wouldn't be surprised if this is what Apple intends to put in some of their machines this spring. I really hope it's something good.
Help I'm a rock.
It seems that they will be going from their current fried potato chip to a more diet efficient baked corn chip.
Will I still be able to run my current x86 OSs and code on these chips?
You are not the customer.
Didn't they make about the same claims when they came up with their 64 bit CPU that was not backware compatable and didn't give up real estate required to run the old legacy code? AMD rubbed their nose in it. Intel, like the customer, is a victim of their own marketing approach over the years. The customer and even the industry has learned to acceept awful penatalties in order to run old outdated legacy code. It's a bad design, but one that was promoted strongly by Intel. Are they doing anything different here to make one believe that this time they will be more sucessful than their last 64 bit cpu that wouldn't run legacy code?
I'm an American. I love this country and the freedoms that we used to have.
The word "multiprocessor" should be "multicore". They're talking about 4 or 8 cores on a single CPU, which might be nice for blades but not so useful for a laptop or a gamer.
And of course, Macheads note the phrase "performance per watt".
They better not be going to PowerPC architecture! *Shakes fist* Darn you, Apple!
to detract attention away from Dell.
Why wait for Intel when AMD already has true dual-core?
The article seems to pretend that the Israeli design teams low power Pentium M doesn't exist. It says the last major design change was the Pentium 4 (which was prior to the Pentium M), and doesn't mention that current and (already announced) future Pentium M based designs match the description given.
Sure the 686 architecture certainly has been around for a while, but the PM is a pretty damn good chip.
It's sad, but the era of exotic CPU's in servers and workstations seems to be ending; X86 is just better "bang for the buck", so much so that even Intel can't compete with it (Itanium)! I hope they know what they're doing.
We apologize for the inconvenience.
Speaking of logos. Did Intel ever use a logo that looked like the current AMD logo?
I have this old 286 motherboard with a soldered-on Intel CPU, but the logo looks like AMD's.
Any ideas?
Oh sweet! That sentence was written so balmily I think it has even qualmed my pre DRM large-scale nationwide deployment fears.
This is what I have been predicting: Apple had a seemingly better choice with AMD's current processors (on a performance per watt basis). However, Intel have already showed their pipeline to Apple and this is what prompted Apple's decision to migrate.
More
Namaste
Intel Corp announced new management roadmap
Because of P IV scandal, board of directors argues changing monkeys with donkeys...
Also in another news, chip maker giant seriously in deepsh*t because of that compatibility of x86 leads them seriously trouble. 25 years ago 8088 was brilliant move and now they cannot create any competitive product because of backward compatibility of x86.
After dumping billons on Ar-Ge they can only create chip named Itanic.
An insider informer says, after seeig AMD opteron sales:
Oh s**t...
[My english is better than most other people's Turkish, so please point out mistakes politely. Thank you.]
Its going to be strikingly similar to Athlon 64.
'Bout time they admitted the P4 burst arch is antiquated.
Raydude
I wonder if they plan to incorporate this technology into their design at all.
Perhaps its still too new, but you'd think they would be looking to the future for ideas...
And they said zombies weren't real!
One thing the article didn't make clear is what exactly Intel means by "A New Chip Architecture".
... it's either a new core or a new spin on the P6 core.
... it's yet another Funkitecture, with a good JIT in firmware on chip, like Crusoe.
... it's one of their existing Funkitectures, ditto.
... it's option 3, with Alpha as the CPU and Freeport Express as the JIT.
... it's option 2-4 with the JIT in the BIOS, like Alpha's HAL.
Option 1 (probability: high)
Option 2 (probability: medium)
Option 3 (probability: medium)
Option 4 (probability: low)
Option 5 (probability: medium-high)
I hope it's not options 2 or 3. Intel has a really bad track record in designing actual processor instruction sets.
Will it run Lotus 1-2-3?
---- "Logoff! That cookie shit makes me nervous!" - A. Soprano
More efficient. More powerful. Great for games too !
If they sold one at the store that had 2 of these chips in them and ran XP/game and linus I would never look back at serial General purpose chips.
http://www.gpgpu.org/
Or at least thats what I think.
The Doormat
If you're not outraged, then you're not paying attention.
Well, Risk and Alpha are going away, and Itanium is the way of the future for HP-UX and OpenVMS. What was interesting was what they told us about the forthcoming Intel processors - the entire Alpha team was hired by Intel and the next gen intel chips will use the Alpha style switchless mesh architecture. This style architecture removes roadblocks inside the box -- no more intermediaries. Your data takes a straight line to its destination.
In other words - it connects processors directly to one another to render true linear scalability. This differs from other architectures in that there is no traditional bus, and you can add processors, memory and I/O capacity in a Lego block-like fashion.
They also mentioned they will be coming out with quad core in 2007.
and instructions. The x86 architecture is due for an overhaul and now is as good a time as ever to do it. CPU speed has stagnated and no new innovations are on the horizon except for the square-peg-round-hole solution of "add more cores!"
I don't want more cores, why is it my Sony PS1 could do all it did with 33MHz and very small RAM? My $20 DVD player can decode MPEG1-4, JPEG, etc. with no massive videocards and CPU's. Sure, these are fairly specialized items, but when you think about it you could take the core of the PS1 and the DVD player and get a computer that could do almost everything a current PC can do for almost nothing cost wise.
x86 has had its day, and Intel and AMD have milked it for all it was worth. It is time to move on and backwards compatability be damned!
http://teasphere.wordpress.com - A little spot of tea
AMD's beating the crap out of us. Let's restart from scratch and copy their model!
Though they may not want to admit it, Intel knows they've lost the 64-bit format war for desktops at least.
So probably what the are working on is a next gen x86 architecture. Those don't come out too often, usually the design one and just modify it for a number of years. It sounds like they are going to start using modifiations on their Pentium M for desktops, which is cool since it is efficient both thermally and in terms of what it does per clock, but there's a limited life to it and they know it. The Pentium M is something of a throwback to the P3, which itself is really based on the Ppro design.
So my guess is Intel figures it's time to unviel a new design for a core, but on x86 architecture.
Considering that no one could make a 64-bit processor as good as the DEC Alpha, and that HP basically GAVE Intel the plans for the Alpha, I believe this new architecture is going to be heavily based on the best 64-bit processor evar: DEC Alpha. It's not so much new as it is just continued development of an already existing superior technology.
-"...bad old ideas look confusingly fresh when they are packaged as technology" - Jaron Lanier (Digital Maoism on Edge.o
"Flyin' in just a sweet place,
Never been known to fail..."
A 64-bit multi-core Yonah-based chip with onboard memory controller and AltiVec (or whatever the generic name is), and HyperTransport.
"That's hot."
You might start here. Lots of other books will tell you how to use semaphores and mutexes. This book will help you to understand why to use semaphores and mutexes (and perhaps open your eyes to better concurrency constructs), and teach you how to reason about your multithreaded design so that you won't get any nasty surprises when it comes time to run it.
And the chips would be cool also...
All processor architectures are based on the algorithm, a custom started by a guy named Babbage some 150 years ago.
A really new architecture should abandon the algorithmic model and adopt a non-algorithmic, signal-based synchronous software model. It would revolutionize computing and solve the nastiest problem in the computer industry: software unreliability.
But we cannot expect a big company like Intel to be truly innovative. Hopefully a bright upstart will get the message and make a killing while the behemoths are busy fighting each other for market share. The won't know what hit them until it's too late.
Just think it will be neat to see if Intel will finaly come out with something that can beat AMD.
I see it interesting that they are doing this right when Duel-Cores are starting to be used and if these are going to be used in Duel-Cores. The less power that it consumes the less heat is the way I see it and Duel-Cores definately need less heat.
he said integrated memory controller, as in hardware. i don't believe intel has a license to this tech. nor do i believe amd would license it to intel cheaply.
sum.zero
exactly what I thought when I read the headline. The timing is too exact. It would have to be somewhat x86 compatable though you would think, else they wouldn't be developing on it now, they'd wait at least for prototype chips.
It's widely known that Intel was going to bring an evolved Pentium M-like architecture to desktop and and server. Better perf/watt. What Apple's gonna buy.
A few hundred "haha see Intel sucks they're stealing ideas from AMD and admiting defeat because they need a new processor OMFGLMAOBBQ"
This is just them building up publicity fo yohan that's it
In your dreams. In most geeks dreams too, because alternative architechtures are cool, and breaking with industry inertia is cool. Even I think it would be cool, but i bet you a fiver it doesnt happen.
use Blunt::Instrument;
It would be interesting to see if that is what the article is referring to.
Processors always used to control memory directly. (Think 68000 or 6502.) This north-bridge gunk is a relatively recent misfeature.
Patrick Doyle
I mod down every jackass who puts his moderation policy in his sig. Oh, wait a sec....
Why is it that nVidia and ATI can make significant changes to their graphics chips every year, but Intel and AMD can only manage seemingly much smaller changes and over a wider timespan?
A core overhaul of Intel's processors should have been done years ago. Intel used to be an innovator. Now they are beginning to choke on AMD's dust.
Kickass Cheap Web Hosting
As long as the gigahertz hype worked, Intel was happy to disregard performance per watt outside of laptops. Only now, when they are hitting clockspeed limits with the P4 and look bad compared to AMD, "performance per watt" is suddenly interesting.
C - the footgun of programming languages
...grab yer ankles!
...they'll go photonic?
At my local PC store, the X2 3800+ is 469$ CAD. The Pentium D 820 2.8Ghz 64-bit dual core is listed as CALL, which means if you have to ask, you can't afford it.
If you split the cost down, you're paying less per core on the X2 3800 than if you bought the equivalent CPUs (which are a bit over 380$ CAD for a regular 3800+).
If they sold one at the store that had 2 of these chips in them and ran XP/game and linus I would never look back at serial General purpose chips
I'm sure he'll be excited his upgrade to this new architecture too!
http://www.gpgpu.org/
Heat and poor efficiency is caused by 'serial' slow general processors. Amd is no different. Scientists are now moving to parallel/streaming processing GPU's. This kind of chip while hard to program for is still superior. The compiler tech is only getting better. Itanium is an excellent chip that you just dogged. Dump the legacy Intel ISA.
The 64bit format war hasn't even started yet. And besides its hard to call it a war when intel doesn't even have a 64 bit chip out. When we see 50%+ penetration of 64bit chips on the desktop then you can start to say who is the winner.
"Now with more flubber!" Yeesh.
So it's good enough for me.
Well, rabid mac fanboy antics aside, I bet that this is what Intel showed Apple to lure them into the switch. In a moment of platform agnosticism, I'll be happy if the end result is that I get a crazy fast and wattage efficient computer in 2006.
Start Running Better Polls
lawyers from HP will come into your office and give your a 2 hour lecture on NDA. When they leave you will be more enlightened. Probably of your soul.
The X2 3800+ has been out for weeks and is currently at $400. It should be down to $300 by year end or less.
[RIAA] says its concern is artists. That's true, in just the sense that a cattle rancher is concerned about its cattle.
.... we used to etch our chips from wood.
You just described the Cell processor architecture. Right down to the number of cores it has.
It is widely believed that this new architechture will be based on, or at least have very close similarities to, the current Pentium M chip design.
Current Pentium 4 CPUs utilize Intel's NetBurst design, which allowed for higher clock frequencies through it's longer piplines. With shoter pipelines, AMD CPUs and the Pentium M have been able to achieve comperable performance at lower clock speeds and at less power.
Less power is the most important part of that, as heat dissipation is becoming more and more of an issue as CPUs become more and more powerful.
Expect the next generation chips to be at a lower clock frequency than the Pentium 4, a departure from Intel's historical patterns.
The new architecture is the Merom family of chips, the replacement for the P4. It is an all new architecture, don't let reports of PM based chips fool you. It is the same philosophy as PM, short pipes, lower clocks, but only tangentially related, and about as ground up new as is possible with modern chips.
If you want more, I put some more details here:
http://www.theinquirer.net/?article=25349
-Charlie
Feel free to mod this offtopic; I was just wondering what the real world advantages of each were, and if asynchronous processors will ever show up in PCs down the road.
its in the subject
Uh... How about, "Wow, I can use my laptop for more than 1 hour at a time!"
They've had the Pentium 4 moniker for longer than either the P3 or P2 I believe. Time to retire it...where's the Pentium 5?
Or...
Why not drop the Pentium name entirely and go with something new? Yeah yeah I know brand recognition...but so what? Intel could just launch a new ad campaign designed around the new name. They've got the bucks and they did it before.
And I'd be really interested in knowing if this new architecture, whatever its called, will be what's in next year's Mactels. (I'm still scratching my head at why they'd drop the bomb then wait an entire year to release any hardware. Seems to me like it'll do nothing but hamper the sales of the PPC line....but I digress...)
Can I hold out a thimbleful of hope that the new chip will be priced sanely? I had wanted to build a mini desktop Pentium M gaming box this year, but it would have easily been a few hundred bucks more than the A64 XPC I built instead.
Why is it that Intel's pricing is consistenly higher than AMD's, yet Intel has the manufacturing capacity advantage?
If you were designing a CPU, how hard would it be to allow the entire state of a core to be duplicated to another core?
My thought is, if on a branch prediction, another core on the chip is idle; copy the state of the current core to the idle one, and each core follows one branch of the code, with one of them being discarded when the branch is resolved. In the case of a resource conflict (FPU, cache memory, or whatever resources are shared between cores) the core running the most likely path gets priority, or in the case of a entirely different thread running at 100% probability, it would get priority over any speculative thread. (but I could imagine that leading to a deadlock, if the 100% thread is waiting for a signal from the braching thread; which is being excluded from the resource by the 100% thread, so perhaps some form of weighted rotation could be used for resource allocation.)
In the case of a long pipeline, a second brach could use up to 4 cores, a third branch 8, and so on, as long as no other threads are utilizing the cores.
If possible, this would add quite a bit of transistors to link the cores together, and manage the states. The advantage would be that single intense threads would be accelerated, because with enough cores, branch prediction failure would drop to near zero (exception being if too many branches need the same scarce resource) while code designed to be multi-threaded would be able to use the extra cores in the ordinary way. The best part is, it would run existing code faster.
1. Both intel and Apple have been saying "power consumption power consumption ..."
2. Apple is busy porting everything they have to x86.
3. I think we can assume Apple would have been given intels roadmap before their recent wedding.
A. 2 + 3 = New chip arch will use x86
instruction set of some type.
4. Simplest way to cut power is to reduce ececution path lengh.
5. Simplest way to cut path length is to cut logic
6. Best shot at loseing logic form x86 is cut things like handleing for different size 16/32 bit registers and dump realmode and a few other things.
B. 4 + 5 + 6 = Smaller faster but not PC
compatible x86 like chip
A + B = New Apple Computer CPU
If apple new in advance which part of the chip would be cut they could have easily managed their porting effors at both the human code level and the compiler code gereration level to make the transition for them almost nonexistant, maybe not even a recompile. Apple feeling like the PPC arch was not competing effectively with the x86s in PCs would have wanted something that was going to let them claim to be faster and better and they would want exclusivity, if not from clones in general at least form clones running windows. So intel without having to improve chip fab technology or come up with anything really new design wise can deliver Apple a faster, less power hungry chip that can operate on a comodity hardware platform. This saves Apple money building the rest of the MAC, provides enough barrier to entry to keep Redmond off the platform long enough for Apple to get it 0\/\/|\|3d, and gives intel a new mainstream market not mostly under Microsofts thumb.
***(I am not suggesting that Intels only market is PC cpus to run windows on, I only state it because it is shrewd on Intel's part, its good to open up new markets no matter what bussiness you are in espicily when someone else has signifigant control to access of one of your other major markets)
Repeal the 17th Amendment TODAY! Also Please Read http://www.gnu.org/philosophy/right-to-read.html
I wonder if they will reverse engineer the Cell CPU
If HP killed the Alpha, explain this press release from Intel dated June 25, 2001, announcing that "Compag will transfer key enterprise processor technology to Intel and consolidate its entire 64-bit server family on the Itanium architecture." Not only was it announced the Alpha processor line was going to be killed, but Compaq management made sure the decision could not be reversed.
According to the press release, there was only going to be one more generation of the the Alpha, and the transition was supposed to be completed by 2004.
Note the bullet point further down in the press release that "Compaq is transferring significant Alpha microprocessor and compiler technology, tools and resources to Intel." As explained further in the press release, "Over the next couple of years, several hundred Compaq microprocessor engineers, compiler experts and infrastructure employees will be offered employment with Intel."
Compaq before the HP merger had agreed with Intel to transfer all of the engineers. There was absolutely nothing HP could have done to have saved the Alpha even if HP had had a sudden conversion and decided to reject the Itanium. HP would have had to have started from scratch with the IP and people already transferred to Intel--it probably would not have been legally possible for HP to have changed direction at that point.
I hope to god English isn't your first langauge,
YES!! I just knew it! Intel is gonna build a PowerPC variant. Steve couldn't let on, becuause, er, well, you know. I know the mighty Apple would prevail. World domination is near. The holy Mac is still the holy Mac. PowerPC is still the future. And now Intel's going to convert! I just know it!
Move along citizens, just having a little delusion here. But hey... rest of hell froze over, I just want a little ice cube, K?
One man's pink plane is another man's blue plane.
This is slightly OT, but I thought I would ask anyway. Does anyone know any field that has benefited from the use of the Itanium architecture? From many of the posts here, it doesn't seem so. I ask because I primarily use computers for computational solvers (e.g. CFD). Recently, HP donated a couple workstations for us to use for research (I'm at a university) and as I was inspecting them I noticed the "Itanium2" badge. After perusing the Fluent (solver of choice) users website, I noticed that they had a version specifically for the Itaniums. Fluent also promised great performance from this platform. While this is nice to know, does anyone here have any experience with the Itanium in an engineering environment and care to discuss? I'm wondering if HP was looking for a way to offload non-selling systems for a tax break (if these are in fact terrible machines, but with 4gb of ram, I guess just about anything can run ok). BTW, they're running Linux.
Cracking a not-as-yet-protected OS to run on a plain vanilla X86 is not the same thing as cracking a protected OS that has been matched up with a TPM and booted from a special BIOS (if the full TCG spec is adhered to). There is a lot of confusion out there about this, but the two are vastly different. While the former is like a "do not enter" sign on an unlocked door, the latter is like a castle with a moat and archers ready to shoot at anything that isn't wearing their colors.
Here's what's really going on. Intel tries to break 4 GHz with the Pentium and fails. AMD "de emphasizes" clock speed with their shifty branding scheme. The chipmakers have hit the wall on clock speed. Either they can't get CPUs going any faster, or it will just cost too much to do so. Intel has more money than God, and clock speed is a proven marketing tool for them even if it's sometimes a misleading metric of performance. If Intel's not making a 10 GHz chip, no one is. It's too expensive.
You remember when computers got twice as fast every 18 months? That's done. Now you'll get this "multi-core" crap which is much harder to program for full efficiency, tends to make memory access into the bottleneck, and thus only improves performance on a small subset of programs which have been specifically crafted for it, and whose algorithms fit the model. Two cores are NOT twices as fast as one. Eight cores are probably not twice as fast as one.
For software developers, the free ride is over. Your downloaded, duct-taped-together, interpreted, scripted code isn't going to magically get fast enough next year any more. Time to learn a compiled language, start with C. Are you an undergrad in a CS curriculum taught entirely in Java? Switch to computer engineering, take a CPU architecture class, and do some embedded coding. Learn how computers work. If your CE curriculum is also all-java, switch schools.
And all you folks saying "practical robot AI is just ten years away" and talking about reaching a "technological singularity"? Don't hold your breath.
What's amazing to me is how the parent got modded redundant, but the 50 posts I saw asking "Did Apple know about this???" didn't.... Wonderful.
Excellent.
AMD already has 4-core Opterons up its sleeve. I'm really looking forward to the next couple of years.
Stick Men