4x4 Chips, Opening AMD's Architecture 229
Nom du Keyboard writes "Once upon a time open slots in a PC that anyone could build a card for were a good idea. PCs with them sold better than PCs without them. Now AMD is proposing another new socket that will be open for plugging in of 3rd party co-processors directly on the processor bus." They've also announced a 4x4 chipset, meant to counter Intel's Core 2 Duo chips. From the article: "Socket 4x4 will have a more immediately impact. Set for a release in the latter half of this year, it essentially lets you combine two dual-core Athlon 64 X2 or Athlon 64 FX chips to create a quad-core desktop PC now ... AMD made the point that Socket 4x4 also provides a more flexible upgrade path for a single motherboard system by letting you start with one chip and add another later on. AMD didn't talk pricing, but you can bet neither the Socket 4x4 motherboards, nor systems that use it to include two dual-core CPUs will be cheap."
4x4 chips! (Score:5, Funny)
Re:4x4 chips! (Score:5, Funny)
Re:4x4 chips! - mod parent funny (Score:3, Informative)
nothing to do with anti-amd sentiments!
(just stating the obvious here, as it seems to be necessary)
Re:4x4 chips! (Score:2)
man those 4x4 cpus are great. what was
Re:4x4 chips! (Score:3, Funny)
*ducks*
Re:4x4 chips! (Score:3, Funny)
4x4? (Score:4, Insightful)
Re:4x4? (Score:3, Interesting)
These systems are designed to handle the dual SLI systems the GFX companies are starting to push.
Re:4x4? (Score:2, Insightful)
Re:4x4? (Score:3, Funny)
10-4
Re:4x4? (Score:2)
Some rough carpentry 2x4s are slightly larger than well-finished commercialized 2x4s. I've seen 1.78x3.82 even. But the idea is the same -- never quite 2x4. The 2x10s I have in my basement at the moment are closer to 1.5x9.6, but I won't have to sand them at all for their intended target project. That's an advantage, not
Re:4x4? (Score:2, Funny)
--
The Earth's biosphere can't be traded in next year like that SUV you're eyeing, dude.
Nice. The dude who's wagging his finger at potential SUV buyers drives a big old truck [oldantiquecars.com] that gets terrible gas mileage (1970-era 5.7L V8, plus 3-speed tranny) and pollutes like a mofo (predates catalytic converters, and due to the age most likely exempt from mandated smog checks).
Re:4x4? (Score:4, Funny)
Re:4x4? (Score:3, Informative)
Re:4x4? (Score:2)
Re:4x4? (Score:2)
Er, no. It refers to a vehicle that has four wheels, all of which are powered. When I was a kid, my dad moonlighted driving a 6x6 towtruck - two powered wheels in front, and four more in the back. It didn't have six wheels all under control of the steering wheel.
Re:4x4? (Score:2)
The Honda Prelude had 4WS
Re:4x4? (Score:2)
Re:4x4? (Score:2)
Re:4x4? (Score:2)
Guns racks? (Score:5, Funny)
While these come with the gun racks standard?
Re:Guns racks? (Score:3, Funny)
Re:Guns racks? (Score:2)
Note, this is a Southern 'merican thang.
Re:Guns racks? (Score:2)
Re:Guns racks? (Score:2, Funny)
Re:A Mod-Est Proposal (Score:2)
Why would anyone help someone who calls people he doesn't even know an asscracker.
Smokey the Bear says... (Score:4, Funny)
Remember what Smokey the Bear says. Only you can prevent your AMD(TM) quad-core desktop computer from starting a forest fire.
Re:Smokey the Bear says... (Score:2, Insightful)
Re:Smokey the Bear says... (Score:2)
Intel responds, naturally... (Score:4, Funny)
JATO? (Score:3)
Sounds neat (Score:4, Interesting)
At my last contract, we used IBM Bladecenters -- Linux in a dev/QA environment, and they had prolly the largest load-generator farm I've ever seen. It wasn't the CPUs that were maxed, tho -- just the network.
Re:Sounds neat (Score:2)
Re:Sounds neat (Score:2)
There again, will this supplant normal quad-processor motherboards? And will they gain quad 4x4 sockets for 16 CPU systems???
Re:Sounds neat (Score:2)
This will reduce by a factor of four the number of expensive low-latency network interconnects needed to build a cluster of a given size.
Not if you want to maintain constant bytes/flop.
Re:Sounds neat (Score:2)
Re:Sounds neat (Score:2)
Re:Sounds neat (Score:2)
Re:Sounds neat (Score:5, Insightful)
I take it you don't do any scientific calculations or physics modeling at your place of work.
And I assume that you don't do 3d animation or video editing either?
Or mabye mass amounts of OCR, Photoshop, or anything else that puts CPU usage at 100%
Sure 90% of the computer market doesn't need this, but the other 10% is willing to shell out the big bucks to be the early adopters. Eventually this will be passed down to the rest of the 90% when the next big thing comes along.
Oh and don't forget the gamers...
Re:Sounds neat (Score:2)
Me, I'll take a 2 CPU/core server with 32 GB of memory over a "4x4" CPU server with say only 4 GB of memory any day.
Re:Sounds neat (Score:2)
Re:Sounds neat (Score:5, Funny)
Re:Sounds neat (Score:2)
Something like this would have been a go
Re:Sounds neat (Score:2)
Re:Sounds neat (Score:2)
Quad machines... (Score:4, Interesting)
Actually if this isn't the case, I'll be very grateful if someone could tell me, because I was thinking of ordering the above for a replacement webserver...
Simon
Re:Quad machines... (Score:4, Informative)
They can; 4x4 appears to be a new marketing label for the same thing. (Just as "Athlon" and "Opteron" are the same chip already.)
Re:Quad machines... (Score:4, Informative)
Has AMD started enabling multiple hypertransport links in the Athlon chips? Opterons have two or three hypertransport links, Athlons only have one link active. Yes, it is artificial, but that makes sure the people that are likely to need it are going to pay for the feature. The multiple links are needed to chain or mesh multiple CPUs together. Maybe the "4x4" chipset is another crosspoint switch to get around the limit of the single link, though it might add latency by adding another hop or two.
Re:Quad machines... (Score:2)
That's what they just announced. What AMD has crippled, they can uncripple whenever they feel like...
Re:Quad machines... (Score:3, Informative)
Tom
Re:Quad machines... (Score:2)
Yes. The new board appears to be more consumer-level and hopefully less expensive.
What I am really hoping for is the vague "other processors" note - perhaps this is also meant as a responce to Cell and the other socket could be populated with a DSP chip. That would be fun !
Re:Quad machines... (Score:2)
You can currently have upto eight dual cores on one network. There are products that extend this with cHT enabled switches so you can go pretty high in the # of nodes.
You can't MP a Athlon64/FX setup because the memory is local to the processor [e.g. not the northbridge] and there are NO cHT links.
Tom
Re:Quad machines... (Score:3, Informative)
In a typical FSB MP system the processors snoop the bus and look for reads/writes.
In the opteron world the processors send out cache probes via the HT links. Athlons have ZERO cHT links activated which means they cannot work in MP systems.
Tom
Re:Quad machines... (Score:2)
Re:Quad machines... (Score:2)
Re:Quad machines... (Score:3, Informative)
Re:Quad machines... (Score:2)
Re:Quad machines... (Score:3, Interesting)
the problem is that they are not like AMD's HyperTransport bus (which makes this really neat) - but wouldn't it be better all around if we moved towards more backplane styles for higher end stuff?
the highest spec backplane i remember was a 64bit 66Mhz PCI bus.. what if we where to move that to PCIe with a massive amount of Lanes.. or have AMD open up their
Re:Quad machines... (Score:2, Informative)
You will need a pretty recent version of Linux. I am running SuSE Linux Enterprise Server (SLES) 9 Update 2, except that I've upgraded the kernel to a 2.6.14 from kernel.org. My suggestion: go with the latest Red Hat Enterprise, or wait for SLES 10, due out any week now.
Wow. (Score:2)
Will AMD hurt itself by undercutting Opteron sales?
Will Intel follow suit with its consumer chips?
Re:Wow. (Score:2)
Its funny that this was mentioned.... (Score:3, Insightful)
Cache coherency? (Score:2)
Given where I work, and that I've never heard of this before today... I suspect it's a hoax.
The only way this would work is if the OS was aware of it and manually routed data from one node to another (e.g. like a northbridge DMA device you can pipe info to).
Tom
Re:Cache coherency? (Score:3, Informative)
Given where I work, and that I've never heard of this before today... I suspect it's a hoax.
The only way this would work is if the OS was aware of it and manually routed data from one node to another (e.g. like a northbridge DMA device you can pipe info to).
AMD's own slides from the 200
Re:Cache coherency? (Score:2)
It could be that they're toying with the idea or the slides could be, oh I dunno, doctored.
Opening the Athlon side up to MP+DC is a really stupid idea because it directly undercuts the Opteron line.
Best I can tell the plan is quad-core on die (as per the CEOs press release), cache improvements, other architectural improvement and the rest filters down into the "consumer" brand. [all public information from press releases]
Tom
Re:Cache coherency? (Score:2)
I just checked my email. It was announced today as a press release by AMD... damn...
They don't really say what "4x4" means... oh well it's not my job anyways.
Tom
Re:Cache coherency? (Score:3, Informative)
Re:Cache coherency? (Score:2)
Know for games to catch up (Score:4, Interesting)
I hear rumors that people use processing power for other things, but I think those are just myths. (Actually I just started to work for a high-performance computing group and they'll probably be excited by the new AMD offerings)
Re:Know for games to catch up (Score:2)
Feature article about it at Tom's Daily (Score:2, Informative)
Yeah, but .... (Score:5, Insightful)
AMD won't happen to produce any of these "3rd party co-processors" will they?
I haven't been this excited since Intel started selling 386SX chips that allowed us
to buy Cyrix (or Intel) math coprocessors for twice what a non-crippled DX cost!
Re:Yeah, but .... (Score:2)
In a way, the 80386SX was to a 80386DX in the same way that an 8088 was to an 8086.
The i486 was the first to have an integrated maths co-processor and early i486sx was
Re:Yeah, but .... (Score:3, Interesting)
No, by definition. If AMD produced them, they wouldn't be "3rd party".
It seems unlikely that AMD would try to get into the coprocessor market. Unless they find an extremely compelling coprocessor idea, they'll make more money using their wafer starts for more Athlon, Opteron, Sempron, and Turion processors than they would by devoting some of those wafer starts to coprocessors.
The example of a security coprocessor is quest
Socket 4x4 will have a more immediately impact. (Score:2, Funny)
Re:Socket 4x4 will have a more immediately impact. (Score:2)
Punctuation, motherfucker! Do you use it?
Re:Socket 4x4 will have a more immediately impact. (Score:2)
So can these 4x4 chips do multiplication? (Score:2)
Now I'll find ET for sure! (Score:2)
Poor Article (Score:3, Insightful)
What does that mean? A motherboard with 2 processor slots? A motherboard that accepts two dual-core processors? We've had both, and for a while.
I wish online editors wouldn't publish meaningless articles like this, and I wish sites wouldn't link to them.
Re:Poor Article (Score:2)
Re:Poor Article (Score:5, Informative)
It's consumer version of a dual-processor Opteron motherboard, with a specific socket layout and memory system that's more directed at consumers. AMD will support this in 2007 with 4x4+ (2 quad-core processors on AM2) and in 2008 with 4x4++, whatever that may be.
These motherboards will also support two x16 PCIe graphics card slots, which if you configured using quad-SLI gives you the other 4. 4 CPU cores, 4 GPU cores.
It's mostly marketing to keep the high end benchmarks in AMD's hands, and thus the kudos, and then further sales.
Quite clear really, although I'm confused as to why AMD didn't go the MCM route on a single socket, like the Pentium D and the upcoming Kentsfield processors from Intel.
Still (Score:2)
So I appluad your information but still find the original article pointless.
Slides from AMD's Presentation (Score:2)
Worth reading IMO.
http://www.aceshardware.com/forums/read_post.jsp?
http://rufus.hackish.org/~rufus/amd/big.html [hackish.org]
I get the feeling that... (Score:2)
This also looks quite a bit like SLI/Crossfire in that it's marketed as "add on more stuff later to boost performance and save money in the long run." It looks nice on paper, but in practice, it's pretty
AMD strategie session (Score:4, Interesting)
AMD has been jacking up their prices which we have assumed is simply a response to their higher quality and increased market share but it has done something interesting. AMD is now selling the majority of high end desktop and workstation cpus but they have low marketshare in high end servers and low end desktops.
It would be easy to claim that these are new strategies implemented by Ruiz their new CEO however they would imply a more stock holder less comsumer driven business and AMD's poor marketing (low marketing budget?) and the $100 laptop project seem to rule out this possibility.
If we look back over the history of AMD it becomes interesting to look at chips like the Athlon MP which went through severe price reductions immediately prior to the release of the Opteron.
Implementing dual cpu chipsets on the desktop is likely a strategie implemented imediately before moving their low end to dual core and their high end to a new cpu architecture.
Amd will likely try to match Intel's price and consumer points, low end desktop (with dual core if my predictions about their consumer centric and engineering company bias are correct), high end desktop (catering to the SLI crowd and consolodating on the likely long term presence of socket AM2 (or subsequent sockets, AMD's 754 for an example of a short term socket), workstation (likely with the same socket but with quad core cpus), low end server (Opteron or replacement) and some kind of new high end chip.
The prediction about a new high end chip is based on reduced gap between the current opteron line and the 4x4 system layout.
All of this is very predictive, but based on my studies of AMD's engineering, ethics, and sales history.
Re:AMD strategie session (Score:2)
No, AMD almost always reduces their prices, just as Intel does. Isn't competition great?
Perhaps what you meant was that the stock market has been raising AMD's share price. That's a different thing entirely. AMD doesn't directly control its share price; that is affected by the market's perception on how well AMD is performing financially, whether they meet their projections and the analysts' forecasts, how their business is perceived to compare with Intel's, and
Re:AMD strategie session (Score:2)
Hardware will be commodity, it's inevitable.
AMD was a good investment when their high end 2600+ was $150 and their fabs were taking off.
What if fab technology takes 20 years to go beyond
Cpu manufacturing will be unimportant, designs like software will push the market and it will move slowly.
If you think it's a good investment go invest.
Won't be cheap? (Score:2)
I remember when people said that about the 80386. They were great for servers, but way too expensive for your own desktop. (Maybe that why I still don't have a 386 on my desktop!)
Today's expensive stuff is tomorrow's obsolete dinosaur. What I'm getting at is this: you can bet it will be cheap. It's just a question of when.
Re:Won't be cheap? (Score:2)
Take SIMD off of the main CPU? (Score:2)
Trying to beat Intel who'll have REAL Quad cores (Score:2)
Re: (Score:2)
CORE ARCH K8, this is shenanigans (Score:2, Insightful)
Re:CORE ARCH K8, this is shenanigans (Score:2)
C//
Re:CORE ARCH K8, this is shenanigans (Score:2)
Re:Good for watching pr0n? (Score:2)
What's that?
You downloaded it?
And you're not in it?
You sir, are a sick individual.
FUD (Score:2)
I have to say, the noise levels in a laptop have nearly nothing to do with the processor used and everything to do with the arrangement of the internal components, airflow, size and speed of fans, etc'. I can think of a few Intel-powered laptops (I own one) that just screams when it's warmed up.
Sure, AMD's mobile stuff is a little less mature than Intel's, but I can say with a fair level of confidence that you h
Re:"4x4 Chips"! COOL! (Score:2)
Are those really your wife's kids names?
(I know they aren't yours, this is SLASHDOT afterall.)
Re:What this means.. (Score:4, Interesting)
This is good as each processor gets dedicated bandwidth which leads to great performance(see Intel's ass-whooping in 4P+ systems for real-world example).
This isn't the case; each CPU has one or two channels of DDR (or DDR2) interfacing with a memory controller on-die. Local RAM is not accessed via Hypertransport. Remote RAM is accessed over a ccHT link, but that's not the same as the non-coherent HT links which are used to connect peripheral buses.
So, what this 1x1 or 2x2 or 4x4 mechanism(and I believe 4x4 is the max it will scale to due to HT addressing limits without external control chips) will allow AMD to do is have 2 cpu's per set of traces to RAM/SB effectively halving the bandwidth that each CPU gets. This would be *really bad* if they were using standard DDR as both those CPU's would be severly starved. But, the fact that AMD has just moved to DDR2, which has a lot more bandwidth than one CPU can consume, should result in a significant net-gain in performance.
I doubt that this is the case. You could conceivably hang a RAM controller off HT in a single processor environment (when the first K8s came out, there was talk of using this as an expansion pathway, although it does awful things to your latency), but if you were to connect some kind of RAM controller to two CPUs at once, how would you do cache coherency? From what I'm reading of this '4x4' stuff, it's most likely that the recent (AM2 and onwards) x2s and FXs ship with a second, cache coherent, HT link enabled (rendering them functionally identical to an Opteron 2xx without the requirement for registered DRAMs.)