AMD Details Upcoming Bulldozer Architecture

Vigile writes "AMD is taking the lid off quite a bit of information on its upcoming CPU architecture known as Bulldozer that is the first complete redesign over current processors. AMD's lineup has been relatively stagnant while Intel continued to innovate with Nehalem and Sandy Bridge (due late this year) and the Bulldozer refresh is badly needed to keep in step. The integrated north bridge, on-die memory controller and large shared L3 cache remain key components from the Athlon/Phenom generation to Bulldozer but AMD is adding features like dual-thread support per core (but with a unique implementation utilizing separate execution units for each thread), support for 256-bit SIMD operations (for upcoming AVX support) all running on GlobalFoundries 32nm SOI process technology."

Comments

Re:Nobody cares.

[Anonymous Coward]

And why, exactly, should/do we not care? This is akin to the announcement of i7 or Sandy Bridge. Maybe if you don't care you shouldn't be reading this story.

Re:Sweeeeet nectar

[Cornelius the Great]

No pricing nor benchmarks. The article is purely a discussion about the architecture.

Not much new information

[kg8484]

Compared to such articles as AnandTech's coverage of this in November 2009, I don't see much new information. Perhaps the key bit, and this is glossed over but you can tell from the slides AMD gave them, is the difference between the bulldozer and bobcat cores. The bulldozer cores contain the two integer units that have been revealed before, but the bobcat core only has one but it still implements hyperthreading.

Re:Not much new information

[Rockoon]

Just to be clear, when you say "integer units" you mean "integer schedulers" and not actual integer execution units, of which even the old Athlon's had 3 per core (and that hasnt changed since then.)

Unlike Intel design, with highly asymmetric execution units, AMD's have had 3 symmetric integer execution units per core since the original Athlons. Its actually a pleasant breeze to write hand-optimized integer code on AMD's.

This new design looks (in the diagram) like it actually has 4 symmetric integer execution units per integer scheduler, with the bulldozer having 2 schedulers per core while the bobcat only having 1 per core (I would guess that the logical cores are alternated on rise-and-fall states of the clock on the bobcat, and the diagram certainly makes it look like that is the case.)

Each seem to have two wide floating point execution units, so the floating point performance of both bulldozers and bobcat's are probably equivalent.

What I think AMD has done here is that with the bulldozer, in integer performance it is going to behave like it has 2x the number of real cores. So an 8 core (16 thread) chip will perform much like an 8 core CPU in floatng point work, but much more like a true 16-core CPU in integer work. This should give it a large advantage over Intel in integer work in equal-core comparisons, but the floating point performance will still lag behind Intel.

Re:Not much new information

[imgod2u]

I believe Bobcat's 2 FPU paths are 64-bits wide. For a total of 128-bits. It initially will not support the 256-bit AVX instructions that are coming with Sandy Bridge and Bulldozer.

Its ALU's also appear to be significantly different than Bulldozer. With only one of the integer units can support multiplies and only two of them can support arithmetics. Two others (using a different scheduler) are load/store units. Bulldozer doubles the ALU resources (but not the number of schedulers) compared to Bobcat. So each scheduler has access to two AGU's, one ADD/DIV ALU and one ADD/Mult.

I was never a big fan of the 3x symmetric ALU's in the Athlons. When it comes to integer intensive code, having a ton of independent ADDs or MULs that I'd need that kind of parallelism for was rare. And the latency (compared to a sane design like Core at least) were significantly higher due to the units being multi-purpose. In either case, with the introduction of SSE2, one could use SIMD if one had a throughput heavy workload anyway.

Bobcat and Bulldozer appears to have moved in the right direction here. I really do like Bulldozer's approach to multi-core and think that with some extension, this could make into very interesting CPU/GPU hybrids as well. Although you could argue it's just another version of SMT similar to Hyperthreading, only with a wider back-end intended specifically for multi-threaded processes.

Re:Not much new information

[Rockoon]

I was never a big fan of the 3x symmetric ALU's in the Athlons. When it comes to integer intensive code, having a ton of independent ADDs or MULs that I'd need that kind of parallelism for was rare. And the latency (compared to a sane design like Core at least) were significantly higher due to the units being multi-purpose.

In the Phenom II design the latency of most of the register to register integer instructions is exactly 1 cycle just like the i7. The units being multi-purpose is not a latency sacrifice at all, although maybe the original Athlons had poor latency for another reason and Agner Fog's reference actually indicates that most of the register to register integer instructions on even the early K7's also had 1 cycle latency.

Even in mem,reg operations, the Phenom II beats out the I7 in latency on many operations (ex: xor [ebx], eax .. 6 cycle latency on i7, 4 cycle latency on Phenom II) .. where the I7 really shines the most is in reading directly from L1 into registers (2 cycle latency vs 3 cycle latency), with a massive 50% advantage on one of the most common operations..

Re:Not much new information

[ravyne]

Actually, from the article at PC perspectives and Anandtech's coverage, I'm under the impression that there are 2x128bit paths in the shared FP unit, and that each path could be scheduled to one thread or another. If that's the case, then traditional floating point and SSE math should perform as if there are two full FPU/SIMD ALUs. Only when using the new, 256bit AVX instructions would you have less performance than a processor with one full AVX unit per processor core. Unless they're saying that the last generation already had dual-128bit paths and could issue two independant SSE operations/clock.

Its really a pretty slick design, and with more and more floating-point work being offloaded to the GPU (and soon integrated on the CPU die) I think this is really going to play out to AMDs advantage. There's been some talk of upcoming Intel processors integrating the 512bit LBNi instructions from Larabee, but I fail to see the point of adding ever-wider vector units to the CPU when the GPU is the obvious place to spend those transistors -- possibly it would be worthwhile if Intel does indeed plan to bring some form of large larabee-core array to the consumer market, either as a GPU or compute card, but those chances dwindle as time goes on. Too bad though, I was really excited about Larabee, from a development perspective, at one time.

Re:Not much new information

[plague911]

That would be an interesting strategy for AMD. Pick one metric and just own the shit out of it. Anything else just midly increase performance. Any users who needs this one metric (in this case integer computation) would be crazy not to go for AMDs cpu over Intel's. This would force Intel into one of three options. 1) Adjust their architecture to match (Unlikely as this takes a lot of time and $$) 2) Cut their prices to even out the battle (Unfortunately they would have to cut their price everywhere which would kill their profit margins on unrelated customers) 3) Let AMD carve out a niche market and let AMD gain significant sales. This seems like it would be a loose loose deal for Intel.

Re:Not much new information

[aiht]

This seems like it would be a loose loose deal for Intel.

And a tight tight deal for AMD?

Re:Not much new information

[dirtyhippie]

What is this about highly asymmetric execution units on Intel? link please ;-)

Anyway, I think you're right about the other assertions. And I like the way that AMD is making a move towards integer performance. It strikes me as a strong move towards the DB-driven, slow dynamic language web serving part of the market. The cost to floating point performance will cost them in the short term in the supercomputing and gaming markets, but they aren't doing real well there anyway on account of Intel's recent strength. I think AMD is betting that their chips with an integrated GPU will be able to make inroads on those markets later this year.

Who needs an FPU anyway? 'G' == 'F'++ ;-)

Re:Not much new information

[Rockoon]

What is this about highly asymmetric execution units on Intel? link please ;-)

Intel Core cores have 6 execution "ports", each serve a range of micro-operations (u-ops) and there is some overlap between them.

Some u-ops can only be sent through a single port, most can be sent through a couple specific ports, and none are suitable for all 6 ports. Most of the integer instructions can only be sent through ports 0, 1, and 5, and these ports also perform some floating point duties. The complexity creates a problem for people tasked with optimizing low level code because they need to be aware of what u-ops are generated by each instruction, and what ports they can be sent to.

This is in contrast to AMD'd setup where most integer instructions break down into u-ops suitable for any of its 3 integer execution units, and that these execution units do not perform any floating point duties.

So optimizing for AMD is a pleasure compared to optimizing for Intel. This doesnt mean that Intel design is stupid or anything, just that its a bitch to hand-optimize for.

The most extensive references arent from Intel or AMD tho, they are from a low level hack named Agner Fog.

Mmm

[elsurexiste]

Call me whatever you want, but the only reason AMD is still alive and well is because they've been innovating and building good products for a while now. Itanium, anyone?

Re:Mmm

[mangu]

the only reason AMD is still alive and well is because they've been innovating and building good products for a while now. Itanium, anyone?

A processor that's nearly ten years old is relevant today exactly how?

TFA says: "all good things must come to an end and with the development of the very impressive Nehalem architecture from Intel, and the upcoming Sandy Bridge, AMDs primary CPU architecture is certainly showing its age"

The market is ruthless, no one buys products from a company that used to do great things if that's all in the past. A new and innovative product line from AMD is long overdue now.

Re:Mmm

[Amouth]

the first AMD64 CPU shipped in April 2003

the first Itanium Shipped June 2001

So the AMD was 23 months late - all they did was tack on to existing x86 where as Intel was trying and did develop a whole new architecture.

almost all of the complaints for the Itanium being slow was due to it having to emulate x86 for software that was not written specifically for the IA64 - Code that was and is written for IA64 runs fast as hell and there is a reason why they are still used today - just is specific applications.

Intel's failure was due to them trying to jump to a whole new computing architecture and expecting programmers to go with them - instead programmers resisted and AMD jumped on that by just extending the existing x86.

Development on what became the IA64 started in 1989 by HP and Intel was brought in in 1994 and the first implementation was in 1998 - hell it is the reason we don't see Alpha's anymore.

AMD64 started in 1999.

So in computing terms AMD had many generations to watch Intel actually Innovate - and then take the short cut to market. Please note I'm not putting AMD down for AMD64, I'm just pointing out that you can not compare the success of it VS the Itanium because they are not the same by a long shot.

Also if you want to learn something new - read up on why IA-64 is so different form x86 and you will see why it is worth investing in. Not for the current project but rather for the knowledge gained by doing it. You would be surprised how much of the R&D that went in to the Itanium is currently running in your newer computers and servers.

Re:Mmm

[petermgreen]

Itanium, anyone?
Yes some time ago intel was screwing arround with itanium (which hardly anyone wanted because it ran x86 code so badly) and netburst (which was slower per clock than a P3) while AMD was pushing ahead with the hammer architecture.

However since core 2 and especially with nahelm (where intel moved to a point-point architecture from a shared FSB architecture) intel has gradually regained the lead starting with the single sockets and gradually moving up to larger platforms. AMD is resorting to throwing cores at the problem in a desperate effort to make up for thier poor performance per core but the trouble is that typical desktop workloads can't really load up four cores, let alone six.

Re:Mmm

[MechaStreisand]

Bullshit. The Itanium is slow as shit because Intel didn't bother to give it out-of-order execution like every other modern processor has. As a result, it is only fast on DSP-like operations and slow at everything else. Out-of-order execution is essential because the compiler can't know at compile time exactly where everything is going to stall: it's provably impossible.

Remember, this is the same company that designed the P4 without a barrel shifter.

Re:Mmm

[imgod2u]

There are plenty of things to learn from Itanium, specifically, what not to do if you want a good general purpose processor. For one, you don't make processor performance so incredibly reliant on instruction scheduling that the biggest compiler team on Earth (Intel's compiler group) couldn't make it run fast on anything except a small subsection of problems.

Secondly, when attempting to gain ISA adoption, making it an exclusive ISA that only you have control and rights to use is a big no no. Sure, it'd be heaven for Intel to be the sole supplier.

And lastly, process and iterations mean more for performance than any fancy ISA. Itanium is consistently one or two process generations behind its x86 counterparts and consistently one or two micro-architectural iterations slower (it takes 2 revisions of the Core micro-arch before Itanium comes out with one).

You can have as clean and fancy of an ISA (which IA-64 was not, btw) as you'd like but implementation matters far far more.

In the end, it wasn't fast enough (the best it ever did was match its x86 counterparts) and it didn't have any other advantages to warrant the switch.

Now, ARM on the other hand....

Re:Mmm

[Anonymous Coward]

It's not that it didn't bother, it used fundamentally different methods. Part of the mistake was assuming that the compiler would pick up more of the slack from the programmer, when we in fact found a movement toward interpreters and JIT compilers that need to have very low overhead, with relatively little advancement in compilers and all that advancement essentially focussed on more established architectures. It might have worked better if Intel themselves came up with a good Itanium compiler quickly.

Re:Mmm

[Antisyzygy]

Its relevant because you can learn from history. AMD will be on top in the future at some point. Besides, why would you want people to stop buying AMD? AMD will eventually go bankrupt without supporters like myself. Then Intel will be the only CPU provider and will end up jacking their prices even higher than they are now and then they will release inferior products.

Re:Mmm

[Nutria]

but Itanium's great for its niche.

Itanium really sucks.

Itanium II is really good at FP. But that's a *tiny* niche, and x86_64 is catching up.

It didn't supplant X86.

But Intel wanted it to. That's why they had to copy AMD's 64-bit architecture to stay relevant.

Re:Mmm

[hedwards]

They've been competing just fine, otherwise they'd be out of business. Intel has been creaming AMD on performance at the high end for some time, but at a cost, the price of the high end chips has been much higher than that of the high end AMD chips. And you really don't get that much more performance for your money. Sure if you really need that performance you're going to pay for it, but most people don't really need that last bit of performance. Still, it is overdue that AMD introduce a complete refresh like this, I just don't disagree that it's been as lopsided as you suggest. Intel has for as long as I can recall had a performance advantage over AMD, it just also pretty much always cost a lot more as well.

Re:Mmm

[coredog64]

Development on what became the IA64 started in 1989 by HP and Intel was brought in in 1994 and the first implementation was in 1998 - hell it is the reason we don't see Alpha's anymore.

The reason we don't see Alpha anymore is that Intel coerced HP to buy up Compaq and kill it off by offering to assist HP in porting HP-UX to Itanic.

Re:Mmm

[Daniel Phillips]

EPIC was a dumb idea.

Re:Mmm

[Aboroth]

Intel didn't give it out-of-order execution capability for a very good reason. They had a lot of faith in their compiler people. Specifically, they expected their compiler teams to perform miracles by performing magic and violating causality by looking at what the processor would be actually asked to do in the future.

Oh wait, did I say "good" reason? I meant "retarded". Yeah, that's the word.

Re:Mmm

[MechaStreisand]

That made me laugh. Thank you...

Re:Mmm

[seeker_1us]

Do you not understand branch prediction and out-of-order execution and how it limits performance? Why do you think it can be done better in hardware than compiler space? Do you think that AMD cpus or other Intel CPUs don't "perform magic and violate causality by looking at what the processor would be actually asked to do in the future?"

Re:Mmm

[G3ckoG33k]

Did you write this? http://emulators.com/docs/pentium_1.htm :)

Re:Mmm

[TeknoHog]

The Itanium is slow as shit because Intel didn't bother to give it out-of-order execution like every other modern processor has.

X86 didn't have OoO until the Pentium Pro, and new x86 models are still being produced without it, for instance Intel Atom and Via C7. So maybe it is a question of implementation, not architecture?

Re:Mmm

[Whalou]

Remember, this is the same company that designed the P4 without a barrel shifter.

This looks like a car analogy.

Re:Mmm

[tuxicle]

You think they'd have learned that after the early '90s i860 fiasco....

Re:Mmm

[AmiMoJo]

Strangely the early Atom CPUs didn't do out-of-order execution either and once again performance sucked because of it. The newer ones do and they feel a whole lot more responsive.

I can't figure out why Intel seems to have problems with OOOE. The low power Athlons which are similar to Atoms for power consumption do it. Well, okay, AMD did create the Turon which was a bit of a disaster, but at least it did OOOE.

AMD's stagnant?

[iamhassi]

"AMD's lineup has been relatively stagnant while Intel continued to innovate with Nehalem and Sandy Bridge (due late this year) and the Bulldozer refresh is badly needed to keep in step."

AMD just came out with Six-Core processors for $200, how is that stagnant? Intel's only 6-core processor is still $1000

Re:AMD's stagnant?

[TrisexualPuppy]

Likely another Intel fanboy trying to spread FUD about the company that he doesn't like and at the same time getting his username posted on the front page.

AMD may not have the resources that Intel does, but it isn't as though Intel is walking AMD around on a leash. This mindset gets annoying after a while.

Re:AMD's stagnant?

[SQL Error]

AMD are selling six-core dual-socket CPUs for $200 now. They're not quite as fast as the Xeon 5500/5600, but the price/performance is awesome.

Re:AMD's stagnant?

[blair1q]

"AMD's lineup has been relatively stagnant while Intel continued to innovate with Nehalem and Sandy Bridge (due late this year) and the Bulldozer refresh is badly needed to keep in step."

Likely another Intel fanboy trying to spread FUD about the company that he doesn't like and at the same time getting his username posted on the front page.

The facts in that quote were presented clearly. AMD is a generation behind on architecture, trying to get comparable performance by multiplying old cores, while Intel has been advancing architecture and multiplying cores at the same time. For about 4 years now, Intel has had 2-4 chips performing at levels above anything AMD could produce.

It remains to be seen if Bulldozer will put AMD anywhere near at-par on a performance/core basis, but it's not 2002 any more, and AMD has no hope of a performance lead.

Re:AMD's stagnant?

[adbge]

AMD just came out with Six-Core processors for $200, how is that stagnant? Intel's only 6-core processor is still $1000

I can't tell if you're trolling or not.

Intel's hexacore offering features hyperthreading technology, which allows each core to execute two threads simultaneously. This means that Intel's hexacore chips actually have twelve logical cores, while the AMD hexacore chips only have six logical cores. Your number of physical cores comparison is meaningless, and actual performance benchmarks show that the Core i7 980X is more than twice as fast as the AMD Phenom II X6 1055T. [1]

[1] http://www.cpubenchmark.net/high_end_cpus.html

Re:AMD's stagnant?

[Anonymous Coward]

Actually, AMD's has a great chance of beating Intel in the future. You fail to recognize that AMD has ATI now and they are going to be fusing CPU's and GPU's onto the same die in the future. They benefit from the experience and IP of ATI. Intels graphics capability so far has been a joke.

Re:AMD's stagnant?

[cynyr]

but not per $, which is the whole point. Sure i can build a screaming rig using a $1500 intel CPU, and a $400motherboard, and then toss in the ECC ram that board needs... and all of a sudden i could have bought a honda civic....

Or i could get 80-90% of that same rig, in certain loads 120-150%, for $500 from AMD.

Re:AMD's stagnant?

[h00manist]

Funny how so many people say competition is great, then turn around and start promoting the strategy of eliminating the competition, supporting whoever is "winning", etc.

Re:AMD's stagnant?

[Tanktalus]

I used to have a P4 with HT until some piece of the machine became unstable at normal operating temps, and then got an i7 quad-core with HT. In between, I also purchased an AMD Phenom quad-core, which is still running, though it probably doesn't have enough fans.

I wouldn't count hyper-threading as doubling the CPUs. Often times I would run a single CPU-bound app and find that the "hyperthread" CPU to also spike to 50-100% as shown on conky. So while you may sometimes see a doubling of your processing power with a hyperthread, my experience says to count a hyperthreaded CPU as if it were a half-CPU.

There's no doubt that my i7 is faster than my Phenom. Then again, there is about a two-year gap in purchasing times, so that's expected, and doesn't speak to which company is producing better hardware. All I know is that I would rate my i7 as if it had 6 non-hyperthreaded CPUs instead of 8.

(Running Gentoo, so I get lots of CPU-bound compiling in to do rough seat-of-the-pants informal comparisons... though now I use distcc to get the Phenom to help the i7 and vice versa.)

Re:AMD's stagnant?

[iamhassi]

"Your number of physical cores comparison is meaningless, and actual performance benchmarks show that the Core i7 980X is more than twice as fast as the AMD Phenom II X6 1055T. [1]"

Trolling much? Comparing a $200 CPU with a $1,000 CPU isn't really fair, is it? Someone shopping for a $200 CPU isn't going to even consider a $1,000 CPU and vice versa. Might as well compare a $100,000 Porsche Turbo to a $20,000 Ford Focus.

Might want to follow your own link and look at the far right column with the prices and compare it to where the CPU ranks compared to price:
http://www.cpubenchmark.net/high_end_cpus.html

Re:AMD's stagnant?

[cynyr]

http://www.anandtech.com/show/3674/amds-sixcore-phenom-ii-x6-1090t-1055t-reviewed/3
some loads the 1090T is better than the 980x....

Intel's hexacore is also ~$900, i can buy a whole computer with a X6 and a 58xx or gtx4xx with 8GB ram for that... and according to your benchmark get ~60% of the performance, but they don't list the test rigs, instruction sets used, etc etc...

Also 2x performance for 4-5x the cost? seems worth it huh?

Re:AMD's stagnant?

[bsdaemonaut]

They're narrowing in more on chip/core design than anything else. Throwing more cores into a package that is otherwise identical is improvement certainly, but not innovation. The differences between say a Pentium Dual Core and a Pentium Core Duo of similar speeds highlights the advantages of improving more than the core count.

Re:AMD's stagnant?

[ArcherB]

Intel's hexacore offering features hyperthreading technology, which allows each core to execute two threads simultaneously. This means that Intel's hexacore chips actually have twelve logical cores, while the AMD hexacore chips only have six logical cores.

I think you may be misunderstanding what hyperthreading is. A processor (or core) can only execute one instruction at a time, hyperthreading or not. All hyperthreading does is allow for two sets of instructions to be queued up, so if one thread (or queue) gets hung up for whatever reason, like waiting over a cache miss, the other instructional thread can proceed, rather than patiently waiting in line.

Think of it as one of those tumbling thingies you have to pass through to get into Six Flags or the subway. It's like that, but hyperthreading has two lines instead of one. If one moron has to stop to find his ticket at the front of the line, the other line may move until he finds it.

Your number of physical cores comparison is meaningless...

Um... no. I believe your "virtual" core comparison is meaningless. I'll take a quad core anything over a dual core hyperthreaded-anything-else any day, thank you. Virtual cores don't mean shit until a thread stalls.

and actual performance benchmarks show that the Core i7 980X is more than twice as fast as the AMD Phenom II X6 1055T. [1]

From the site you linked:

Intel Core i7 980X @ 3.33GHz: Score of 10,325 at $989.99*
AMD Phenom II X6 1055T: Score of 5,146 at $194.99*

Hmmmm... Twice the performance at over 5x the cost. Strange, I don't know why you chose that AMD chip. It's odd that you would choose the fastest Intel chip and a middle of the road AMD Chip. Why not this one?
AMD Phenom II X6 1090T: Score of 6,057 at $289.99*.

Oh, I know. Thenyou wouldn't be able to use the 2x faster line. I get it now.

Here, take a look at THIS chart and pay attention to the price/performace graph. You'll see that your chip performs about 2.5x less than the AMD Phenom II X5 965 when price is a consideration. Oh, and for nearly everyone that is not living off their mommy's credit cards, price is a consideration.

Re:AMD's stagnant?

[0123456]

Actually, AMD's has a great chance of beating Intel in the future. You fail to recognize that AMD has ATI now and they are going to be fusing CPU's and GPU's onto the same die in the future.

If you want fast graphics then you buy a discrete graphics card. If you're using integrated graphics you don't much care whether it's a crappy ATI chip or a crappy Intel chip because it won't run modern games at any reasonable speed either way.

Re:AMD's stagnant?

[0123456]

This means AMD old genration processors are capable of the same performance as brand new Intel ones. I don't think this is called being left behind for AMD.

Being faster than a brand new Intel Atom isn't really a great selling point for a modern CPU.

Re:AMD's stagnant?

[MBGMorden]

If you want fast graphics then you buy a discrete graphics card. If you're using integrated graphics you don't much care whether it's a crappy ATI chip or a crappy Intel chip because it won't run modern games at any reasonable speed either way.

That's conventional wisdom, but conventional wisdom doesn't always hold steady in the computing market. 15 years ago what you said there was true for both audio chips and network cards. Anybody who wanted one that was half-way decent bought a discrete unit because those performed well, and the hokey versions that you might find integrated were pretty much junk.

Today? All but a few holdouts and professional level users just use the integrated network and sound, because for your average user - even your average power user - the integrated stuff is plenty good enough.

I'd wager that in less than 8 years your statement of "If you want fast graphics then you buy a discrete graphics card." will sound just as outdated and clueless as "If you want to crunch numbers faster than you buy a dedicated math co-processor.".

Re:AMD's stagnant?

[0123456]

Logical cores are a marketing gimmick. Just like you. See what I did thar?

No they're not. They're particularly important on Atoms, which can't use out-of-order execution to hide pipeline delays and need something to fill up those clock cycles, and the Nehalem architecture has additional execution units to greatly increase the chance that it will be able to execute instructions from two threads in a single clock cycle.

Re:AMD's stagnant?

[petermgreen]

AMD just came out with Six-Core processors for $200 [slashdot.org], how is that stagnant? Intel's only 6-core processor is still $1000 [google.com]
And WHY are they selling 6-core processors so cheap? it's because their quad core processors can't keep up with intels so they are trying to make up for poor performance per core with higher core count.

Trouble is for desktop applications going from 4-core to 6-core is only of marginal benefit.

Re:AMD's stagnant?

[0123456]

I'd wager that in less than 8 years your statement of "If you want fast graphics then you buy a discrete graphics card." will sound just as outdated and clueless as "If you want to crunch numbers faster than you buy a dedicated math co-processor.".

Except there's an infinite capacity to use graphics power, so there's no way that in only eight years we will have reached an effective limit on processing power.

Re:AMD's stagnant?

[imgod2u]

The quote you were fanboy-blasting said nothing about price. Technically minded people are concerned about technical features, such as performance and power consumption. And in that, the quote is exactly right. AMD is a generation behind.

I fail to see what the pricing structure for consumer sales has to do with that.

Re:AMD's stagnant?

[Anonymous Coward]

What are you talking about? Instructions per clock cycle doesn't mean threads per clock cycle. Even old AMDs had multiple instructions per clock cycle, but so what?

Hyperthreaded core != 2 cores, *ever*. Hyperthreading is a way of trying to mitigate stalls. Nothing else.

Re:AMD's stagnant?

[blair1q]

That is how AMD stays in business, by cutting its prices well below the average market price for the performance rating.

But there is a large chunk of performance rating they can't even approach.

Here's last year's numbers (didn't see this year's in the first page of google results), which should give you an indication of why AMD went looking for more performance from each chip. I'm still not expecting Bulldozer to get AMD up to the top. They might match the second- and third-place chips from Intel, but they haven't come up with anything that can outright beat what Intel has fielded.

And AMD's pricing policy costs it a lot of money, since its production costs are much higher than Intel's. They've never had the kind of production efficiency Intel had, and now they have sold off their fabs and are contracting their parts out to GlobalFoundries, who charge a cash profit on each one and still aren't as efficient as Intel.

Re:AMD's stagnant?

[blair1q]

Thinking like that cost AMD $3 billion in goodwill value from when they bought ATI, and led them to have to sell off their production facilities and become a design and marketing company.

"Fusion," the project they had in mind when they made the acquisition, was supposed to be out three years ago.

What they'll release this year or next will be a small, low-performing CPU melded clumsily to a small, low-performing GPU.

And Intel already did it.

Re:AMD's stagnant?

[blair1q]

No, it will always be true.

The amount of stuff you can cram on a single chip is smaller than the amount of stuff you can cram on two chips, and chips that are twice as big are twice as likely to have catastrophic production flaws.

At the consumer level, separate CPU and GPU will be the only way to make a buck, at least until someone reformulates both computing and graphics to be indistinguishable (what Intel was trying to do with Larabee).

Re:AMD's stagnant?

[Mr. DOS]

One thing you've got to keep in mind that more cores, logical or otherwise, only help when you're running multiple processes or processes that spawn (and make effective use of) child threads. If you're doing comparisons with software that's not multithreaded, the only difference you'll see is that of clock speeds and processor efficiency. What I'm trying to say is that you can't expect a eight logical cores to be that much faster than four when the software's only running on one and the difference you're seeing is because the i7 is a better-designed processor.

Re:AMD's stagnant?

[MBGMorden]

There's an infinite capacity to use floating point arithmatic too, but we abandoned the separate chip for it idea long ago. FPU's these days are still getting faster with each chip - no limit on processing power was hit. We simply got to a point where a completely capable FPU could be bundled in with the CPU and it's performance was sufficient for most users.

Imagine this scenario: the integrated solutions don't suck. Instead of being virtually useless for 3D graphics, they have performance about equal to the mid-line $150 to $200-ish cards of today (and let that scale for whatever cards meet that definition of the time). You can get better performance, but it's going to take huge full-length cards running SLI or the like, and it's going to take several hundred dollars to beat your standard integrated solution.

My wager is that 95% of the people who currently buy discrete chips would accept integrated at that point. The chips would still get faster over time, and there still might be a few extreme solutions available, but the average user wouldn't need them anymore. My guess is we'll get there quite soon. And if you're asking why the chip companies would want to sell us 1 chip where they previously sold 2? Simple answer: market competition. If AMD can push out a chip as fast or faster than Intels that also has an integrated GPU that rivals discrete solutions, then they'll take a lot of business from Intel. That's all the motive they need.

Re:AMD's stagnant?

[bhima]

Outside of the benchmark you have listed, how much time does the average business user really use those 12 logical cores? All other things being equal (including price) wouldn't most folks be better off buying few but faster coares?

Re:AMD's stagnant?

[BlackSnake112]

Depends on what you are doing. If you are running a VM (or a few VMs) having the extra cores and a lot of RAM is a good thing. If all your doing is email, basic office tasks, or a few games* maybe not.

*Some of the newer games use all the cores. This may not be true going forward. The latest games may benefit from many cores.

I would say the 'base' new computer should be quad core today. That is overkill for most regular tasks. But it is enough for the rest of regular user tasks. Ultra high end people usually spec out their own system.

Re:AMD's stagnant?

[Tanktalus]

I think you may have misinterpreted what I said.

A single thread of high CPU usage should only impact a single (virtual?) CPU. However, since the secondary (hyper-thread) CPU also was impacted, it tells me that there are some situations where a HT CPU cannot do two things literally at once. For lack of better statistical methods, I estimate this as the virtual CPU counting only as half a CPU.

When I run "make -j13", there are so many processes flying around that I can't tell quite so directly whether the hyper-thread CPUs count as nothing, full, half, or any other portion. Especially when a bunch of the compile processes happen on another physical machine. Determining the value of the extra virtual CPUs here would take far more analysis than I've attempted to do, which I freely admit. (More than I care to do, either - I've already purchased the CPU, and did so based on counting the HT CPUs as half CPUs, for a total of ~6CPUs instead of the 8CPUs Intel keeps trying to call it. Having already purchased it, I want to use it, not analyse it.)

Re:AMD's stagnant?

[Nadaka]

Not always the case. Intels on chip and motherboard integrated graphics are horrible.

But ATI's HD 3300 motherboard integrated line of CPU's will actually run games up to about 2 years old fairly well.

You will always be able to get better discrete GPU's, but it is no longer the case that integrated can not be used for gaming at all unless you need the latest and greatest.

Re:AMD's stagnant?

[ElectricTurtle]

If your logic reflected to any degree the way things actually are, multicore CPUs wouldn't exist. As with CPU production today, dies with bad CPUs or GPUs will just go in the next bin down. Bad CPU? Put it in a discrete graphics card. Bad GPU? Make it a CPU only. Just as bad hexas become quads and bad quads become triples etc. Intel and AMD both do this all the time.

Re:AMD's stagnant?

[ElectricTurtle]

Wow, so you have discovered that the human eye is somehow infinite in its capacity to perceive?! You'll win the Nobel prize for this!

The reason that audio leveled off is that the human ear and capacity to hear is finite. Once graphics are fully photorealistic, there really is no higher level.

Re:AMD's stagnant?

[Ephemeriis]

I'd wager that in less than 8 years your statement of "If you want fast graphics then you buy a discrete graphics card." will sound just as outdated and clueless as "If you want to crunch numbers faster than you buy a dedicated math co-processor.".

Except there's an infinite capacity to use graphics power, so there's no way that in only eight years we will have reached an effective limit on processing power.

Ridiculous.

Once upon a time you bought a machine with a separate, discrete math coprocessor because the CPU couldn't handle math on its own. I remember there being a noticeable difference just running a spreadsheet on a two computers that were identical except for the presence of a math coprocessor. I remember some pieces of software simply refusing to run because I didn't have an FPU on my machine.

Those days are simply gone. These days nobody has a discrete math coprocessor. Sure, yes, you can get fancy cryptographic cards and things like that if you're doing specialized work... But those things don't go into your average workstation. Even heavy gamers and power users don't have discrete math coprocessors anymore.

Something similar will happen with graphics. It is only a matter of time. And I personally suspect that 8 years will be more than enough time to see it happen.

I'm not sure what you mean by an "infinite capacity to use graphics power"... Sure, you can make a scene arbitrarily complex. Keep tossing more shaders and polygons at the screen until the thing begs for mercy. But there's also an infinite capacity to use computation power. I can keep tossing calculations at my CPU until it begs for mercy. In fact, that's pretty much what your average benchmark is.

But that theoretically infinite capacity to use computation power doesn't mean a thing to your average user - or even gamers or power users. Sure, there's still a push for more raw horsepower... But nobody cares specifically about FLOPS on home computers.

And eventually we'll get to the point where the integrated graphics are more than capable of making a very pretty picture on the screen. More than equal to the best discrete card we have today. And it just won't make sense to have a separate card in the machine. Yes, there'll probably be cards available... But they'll be aimed at folks doing realtime photorealistic 3D modeling - not home users playing games.

Re:AMD's stagnant?

[blair1q]

Not so much for you with the "logic" taunts.

You can do things with two simple CPUs better than one complicated CPU. And you still have a GPU to do the video.

Putting a simple GPU and a simple CPU on one chip is not the same as having a complicated CPU and complicated GPU on different chips, and putting a complicated CPU and GPU on a chip is way more expensive than putting simple ones together.

The "logic" isn't immutable with price point. CPU/GPU combinations will be the way to go in palmtop/smartphone devices at low prices, but their performance will be pathetic compared with discrete chips in desktop or laptop computers.

Re:AMD's stagnant?

[Vancorps]

While AMD inarguably has been a manufacturing generation or two behind for quite some time you fail to realize that the platform was design from the ground up to be SMP friendly. This has helped AMD pretty much rule the roost for several years in the virtual market. The 12 Core Opterons are beating the hell out of the Xeons on price and performance. Intel made a lot of great strides to improve their situation but in the end AMD has been pretty good about maintaining their lead there. AMD also has the additional benefit of maintaining their sockets for longer so many older AMD servers support the latest 12 core Opterons. HP and Dell like to lag their bios support to encourage you to replace the whole server but those building based on Supermicro or any of the custom built server markets can do in-place upgrades adding significant capacity to the datacenter for comparatively little upgrade cost.

Intel rules the performance desktop market but AMD has the virtual market which IMHO is the market that has a future.

Re:AMD's stagnant?

[hairyfeet]

Thanks for that link. The "bang for the buck" along with putting my money where my mouth is after finding out Intel rigged the game with backroom payoff is the reasons (well, along with Nvidia and their bumpgate bullshit) why I switched my shop to building AMDs exclusively. For the average person, even power users like myself, the AMD triples and quads are frankly overkill.

My personal PC, which was paid for by profits selling AMDs (thanks AMD!) has a 925 quad, 8Gb of DDR 2 800Mhz, an HD4650 with 1Gb (not big on gaming but works GREAT for transcoding!) 2 500Gb HDDs, and Windows 7 HP x64. The grand total? $659 before MIRs and $575 after. Bioshock 2, transcoding videos, multitasking the hell out of the thing, and frankly it purrs like a kitten. And the bang for the buck is just unreal from AMD, so I frankly don't care if their arch is a little long in the tooth or not. Hell the prices were so cheap I went ahead and built my dad a quad, will he ever need that much power? Probably not, but he can run video chat while surfing and having that bloated piggy Quickbooks running in the background without a hiccup, and at an extra $40 why the hell not go quad?

You really have to look at both the price/performance and more importantly what the average person does with a PC. If you do you'll see Intel is simply waaay overpriced. Most folks would never be able to actually tell a difference between even an AMD triple and the new i7s, simply because most of the time the PC is waiting for them to do something. For videos, surfing, office work, mainstream gaming, you really can't beat AMD. Hell for the price of a single loaded i7 980x I was able to deliver TWO AMD quads to the print shop down the street last week, with 4Gb of RAM each and HD4650s with 1Gb so they'd have plenty of video RAM for buffering. Talking to them this morning they couldn't be happier with the performance, and want me to work up a proposal on outfitting the rest of the office with new AMDs. I'd hate to even think what that would cost if I had went with i7s.

Re:AMD's stagnant?

[ElectricTurtle]

I would be interested in a real world contrast of supposedly 'simple' vs. 'complicated' CPUs. Current multicores are designed to scale. That's why failed multicores are able to downscale. A Core Solo is not a 'complicated' single CPU, nor is a Core Duo a pair of 'simple' CPUs. They are virtually identical, just scaled and linked. The same is true of every other recent architecture.

Your hypotheticals about cost are only so much blather until the market ultimately decides what works at which prices.

Re:AMD's stagnant?

[imgod2u]

No, what happened was that the most FP intensive tasks (rendering, 3D modeling for games) moved to another dedicated chip (the GPU). Bigger and better compute capacity there has not stopped being in demand ever since and shows no signs of slowing down.

The only thing left that were really compute intensive on the CPU were things like video transcoding and precise (production quality) 3D rendering due to the lack of double-precision support in GPU's as well as the difficulty of using them for compute (i.e. writing back to memory) purposes. For the vast majority of common consumer tasks, there really wasn't much demand for ever beefier FPU's.

Also, just as what may eventually be true for GPU's, there are advantages performance-wise of moving the FPU closer to the CPU. In the case of the FPU, it was much easier to tie in FP intensive tasks with control structures the CPU provided. I would suspect that in the case of the GPU, it will be removing the latency of the PCI bus and the need to copy memory between main and GPU memory, potentially allowing a lot more interaction between the physics and AI with the rendering process.

Re:AMD's stagnant?

[Runefox]

Except that fully photo-realistic graphics have been a long time coming, and will continue to be. We're still using the same resolutions we used a decade ago, we just have better models and fancy post-processing effects that make it look that much better. Strip away the pixel shaders, lighting and texture filtering, and the models and textures are still actually pretty ugly today. We can start talking photorealism when we begin to see screen resolutions in the vicinity of 4096px wide on 17" displays and texture resolutions far beyond that. By then, I'm guessing we'll be using realtime ray tracing instead of the traditional styles of raster graphics we use today.

It's also likely that by that point, yes, CPU's would be capable of handling that computational task fairly well, since ray tracing performance is as I understand it fairly linear and independent of texture resolution or things of that nature. We'll probably first see dedicated GPU's for the task, but then watch it evolve the same way we have with audio and other chipset features. Truth be told, that would simplify the life of the PC gamer by a wide margin.

Re:AMD's stagnant?

[ArcherB]

A processor (or core) can only execute one instruction at a time, hyperthreading or not.

Uh, no. That hasn't been true for years.

Nehalem, I believe, can execute up to five instructions per clock per core; though you'll rarely be able to reach that limit.

I said one instruction at a time, not per clock. Still, the point was that when something went awry, the entire instruction chain would have to be dumped and reloaded. The processor would sit idle while waiting for this to complete. The P4 had an extremely long instruction pipeline that allowed it to clock at higher speeds, but choke when pipeline had to be flushed. This is why the Athlons of the day were much faster at lower clock speeds due to the shorter pipelines. Whenever a stall happened, the Athlon didn't have to sit idle for nearly as long as the P4.

Hyperthreading was patch for this. It was nothing more than a second instruction pipeline. When an instruction pipeline was flushed, rather than sitting idle, the CPU would run from just the secondary pipeline until the original pipeline was filled.

Regardless of the number of pipelines, the execution unit was not duplicated, only the instruction queue. I don't care if a processor can perform 5 instructions per clock or 1000, when it stalls, it performs exactly 0 instructions per clock until the pipeline is filled again. Remember, all processors wait at the same speed.

Wikipedia has a pretty good write up on this.

Re:AMD's stagnant?

[Bengie]

the real question is how does Intel vs AMD compete a price ranges. Who cares that Intel's top end CPU is $1k+. I can get an i7 and OC it to 4ghz and beat the crap out of AMD.. at least for now. and still keep a close cost

CPUs tend to be better at different things. From the sound of it, AMD is claiming about a 50% increase in performance because of per-core reduced die space while offering a relative per-core performance.

How will that compare to Intel's 32nm 6core with HT?.. who knows. I hope AMD does well because we need the competition.

Can't wait for 2011. The year of SSDs and many core CPUs

Re:AMD's stagnant?

[bhcompy]

Hybrid Crossfire begs to differ. Build an integrated architecture and those that still want discrete graphics cards will still benefit while the reduction in sku's improves production efficiency. Hybrid Crossfire, of course, being another innovation that AMD has taken to the market ahead of anyone else despite limited resources. It's not so much about engineering new architecture.. you can't do that with those kind of resources all of the time. Instead, innovate existing architecture, which is cheaper to develop, takes less time, and is better for the consumer and product developers. Doing things like increasing the instructions per clock cycle during the mhz wars are reasons that AMD still exists while other companies like Cyrix are afterthoughts. Innovation doesn't exclude extreme optimization.

Re:AMD's stagnant?

[Belial6]

For the most part, I feel like we already are there. Out of the 9 computers running in my home, only 1 has a discreet graphics card, and I am toying with the idea of removing it because it increases the watts used by the computer over the integrated solution. While we don't play a lot of games that are just released, there are at least 40 man hours of PC games played in my house each week, so it isn't like we don't play games.

Re:AMD's stagnant?

[Bengie]

My cousin was doing server benchmarks before doing a large purchase for his datacenter and he found the i7 with HT disabled was still beating the AMD and with HT it about doubled in speed. He runs a mix of DB/Video Compression and everything is ran on Solaris/Linux and stores several petabytes of data.

I guess consumer grade apps/hardware tend not to enjoy the extra kernel threads so much.

Re:AMD's stagnant?

[Lally Singh]

Yeah, you get much better performance going through a northbridge & PCI-e than talking to a gpu on the same chip...

Re:AMD's stagnant?

[ElectricTurtle]

Actually we're using the same resolutions mainly because displays are the limiter. All the super awesome cards at the top of the market right now can render resolutions twice or more what most people can display because their monitor sucks.

Re:AMD's stagnant?

[bhtooefr]

No, it will always be true.

The amount of stuff you can cram on a single chip is smaller than the amount of stuff you can cram on two chips, and chips that are twice as big are twice as likely to have catastrophic production flaws.

At the consumer level, separate CPU and FPU will be the only way to make a buck, at least until someone reformulates both computing and floating point to be indistinguishable (what Intel was trying to do with 80486).

Actually, the comparison is more apt than it sounds, even - a GPU is basically a ridiculously parallel FPU. And Nvidia and ATI are both working on pushing that "ridiculously parallel FPU" part, reformulating computing and graphics to be indistinguishable.

Re:AMD's stagnant?

[IICV]

Even better - imagine what this would do for laptops! No more would you have to toss out your old one because the GPU has some weird proprietary socket for which nobody's ever produced another chip, so even if you upgrade the CPU that'll still be a bottleneck - you just swap out the GPU/CPU package, and you've upgraded.

And since the GPU and CPU are sharing the same physical space, you can use more case area for things like cooling and memory and disk drives and what have you. It'll be awesome.

(yes, I'm assuming that this is the future - after all, isn't a graphics card basically equivalent to a multi-core CPU, except optimized for different operations?)

Re:AMD's stagnant?

[Kjella]

Somehow I doubt you can put mid-range cards that still draw 50-100W together with the CPU without creating a very complicated heating solution. Integrating everything is more a cost saving for the low end not to mention a way to sell a graphic chip with every machine keeping the competition out. I'd say the threat here is actually much greater the other way around, after all Intel is not hurt much directly if all AMD chips are sold with AMD graphics - that is nVidia. But if Intel can improve their integrated designs or squeeze nVidia into licensing them some then AMD would lose a huge market. In short it boils down to that you can add a discrete graphics card if you're unhappy with Intel's GPU performance, but you can't add a discrete CPU if you're unhappy with AMD's CPU performance.

Re:AMD's stagnant?

[TemporalBeing]

"AMD's lineup has been relatively stagnant while Intel continued to innovate with Nehalem and Sandy Bridge (due late this year) and the Bulldozer refresh is badly needed to keep in step."

Likely another Intel fanboy trying to spread FUD about the company that he doesn't like and at the same time getting his username posted on the front page.

The facts in that quote were presented clearly. AMD is a generation behind on architecture, trying to get comparable performance by multiplying old cores, while Intel has been advancing architecture and multiplying cores at the same time. For about 4 years now, Intel has had 2-4 chips performing at levels above anything AMD could produce.

It remains to be seen if Bulldozer will put AMD anywhere near at-par on a performance/core basis, but it's not 2002 any more, and AMD has no hope of a performance lead.

It's more the other way around. AMD has been using the AMD64/Athlon/Opteron architecture for a while now, yes. However, in that time Intel has (i) tried to move to Itanium and failed, (ii) tried to replace the Pentium Pro architecture with Netburst and failed - ultimately reverting back to the Pentium Pro architecture that is still at the heart of their latest processors, (iii) tried to compete with AMD only to adopt most of their general designs (e.g. memory controller on-chip) and their 64-bit instruction set, and (iv) developed the Core series processors. Meanwhile, AMD has been doing a fine job competing on all fronts with the same architecture, and minimal extensions to it.

Re:AMD's stagnant?

[blair1q]

Since Core i7 came out, the Pentium core used in the Atom became simple.

The newer core is usually more complicated and can accomplish more in less time. This is the application of Moore's law. More transistors means more complexity including things like parallelism and pipelining and lookahead and cache and so on. Cores get progressively more complex, and get combined into progressively more complex devices. Intel is planning not just CPU/GPU hybrids, but CPU+GPU+security+communication+yadda+blinken+nod hybrids. You name it, they can drop a special-purpose core in for it, and the number of cores is essentially unlimited as long as Moore's law keeps working.

But for all of those, you can make a collection of chips with discrete cores of various types, each of which is much more complex than those in the hybrid, and it will perform better overall, even if it's a little slower on the few things that depend primarily on interconnect for speed.

Re:AMD's stagnant?

[0123456]

What are you talking about?

I know what I'm talking about. Apparently you don't.

Instructions per clock cycle doesn't mean threads per clock cycle.

Nehalem can, as I understand it, execute up to five instructions from two threads per clock cycle on each core, if sufficient resources are available, based on Intel documentation and numerous articles on its architecture.

If you believe otherwise, perhaps you could provide some proof?

Hyperthreading is a way of trying to mitigate stalls. Nothing else.

Not true. So I don't understand why you've been modded 'insightful'.

Re:AMD's stagnant?

[ShakaUVM]

>>Something similar will happen with graphics. It is only a matter of time. And I personally suspect that 8 years will be more than enough time to see it happen.

Doubtful. Look at the transistors on a CPU vs. a GPU, how they're allocated, and projections for transistor budgets in 8 years.

Re:AMD's stagnant?

[aiht]

I'll chip in with my anecdotal data point:
I'm using the integrated Radeon on my home PC, and I have yet to try a game that doesn't run well on it - Left4Dead 2, Unreal III, Starcraft II... not the most cutting edge graphics, maybe, but hardly ancient.
The only reason I might get a discrete video card will be for things like TV out and better multi-monitor support.

Re:AMD's stagnant?

[ArcherB]

From your quote:

The technology enables a single physical processor to execute two or more separate code streams (threads) concurrently using shared execution resources. ...

Each "execution resource" can only execute one instruction at a time, period, always and forever. What I am calling "instruction pipeline", Intel is calling "code stream". Their write up is careful to say "execute two or more separate code streams (threads) concurrently using shared execution resources". Notice the "execute two or more code streams", not "execute instructions". It also specifically points out "... using shared execution resources", meaning that each "code stream", or instructional pipeline, has to share the execution resource.

Think of it as two lanes merging on a highway. While both lanes move forward at the same time, only one car gets through at a time. If a car in one lane stalls, the other lane can still move freely until the stalled car is restarted. Only in computers, if you get a stall, everyone in that "lane" has to go home and start over.

Re:AMD's stagnant?

[GooberToo]

Intel's hexacore offering features hyperthreading technology, which allows each core to execute two threads simultaneously. This means that Intel's hexacore chips actually have twelve logical cores, while the AMD hexacore chips only have six logical cores.

Virtual core count is a completely meaningless metric.

You do realize that simply enabling hyperthreading can actually reduce performance right? In some cases significantly so. The general rule is, unless your primary applications are specifically coded and compiled to properly support hyperthreading, you're going to take a performance hit. Most benchmarks (admitted, dated as its been a while since I last looked) come in around a 30% performance LOSS! That's based on Intel CPUs. We have no idea if AMD's implementation will suffer the same constraints. If not, AMD could potentially have a huge performance gain on a broader, less specialized workload.

Not to mention, proper hyperthreading support requires proper OS scheduling support. Last I looked, Window's scheduler provided less than ideal support for many work loads while Linux tended to perform almost ideal.

Unless you know for a fact your applications specifically take advantage of hyperthreading, enabling hyperthreading is likely SLOWING you down; and potentially, even more so on Windows. And this completely ignores the fact those almost all of those published benchmarks are for dummies who have no idea what they really mean. If you really want to know how a CPU is going to perform for your workload, you'll almost always have to benchmark it yourself - and the results almost universally look nothing like the widely available benchmarks.

So to bring this back on topic, your argument is that Intel is faster because they are more likely to inflict a significant performance penalty unless you know for a fact your applications will benefit. Hmm....seems dubious at best.

Re:AMD's stagnant?

[GooberToo]

They absolutely are not a marketing gimmick. They absolutely can lead to significant performance loss. In fact, if you have hyperthreading enabled, you're likely running slower because of it! The vast majority of people should have hyperthreading disabled - unless you just like running slower than possible.

Re:AMD's stagnant?

[GooberToo]

A single thread of high CPU usage should only impact a single (virtual?) CPU. However, since the secondary (hyper-thread) CPU also was impacted, it tells me that there are some situations where a HT CPU cannot do two things literally at once. For lack of better statistical methods, I estimate this as the virtual CPU counting only as half a CPU.

Then you are being generous. There are a lot of misconceptions of how and what hyperthreading does. Hyperthreading is basically a CPU mechanism is lowers the cost of performing a context switch from one thread to another so long as the thread is executing on the same core. This is why OS scheduler support is so important and why a scheduler mistake can completely blow any and all hyperthreading benefits for the given scheduler interval. Additionally, hyperthreading attempts to schedule units of work on the single core when the core otherwise has scheduling gaps available on various execution units. To gain full benefit, you need an optimizing compiler which understands hyperthreading. Mostly this means Intel's compiler. Basically you need a unit of work to not require scheduling on all available execution units and the hyperthreaded unit of work must be able to be executed on the idle execution unit.

What this means is, if a core is fully utilized or the available instructions are not laid out in a manner which allows the hyperthreading scheduler to effectively schedule units of work, there is no performance gain. In fact, because of the extra context switches and resources which are still used attempting to schedule work which can not be scheduled, a performance penalty of roughly 30% and higher latencies is very common. Add a little more if not running on Linux.

In the end though, even when everything is working correctly and your applications are specifically targeting hyperthreading support, chances are you won't see more than roughly a 30% gain - because ultimately, you still only using spare execution unit capacity; which generally isn't a lot.

Most people, by far, unless they specifically know their applications take advantage of hyperthreading, perform better with hyperthreading disabled. In fact, server installations are always recommended to disable hyperthreading, again, unless you specifically know your server applications directly benefit from the use of hyperthreading. I consider this to be the desktop rule of thumb too, again, unless you know. An example would be Photoshop being your primary application.

So at the end of the day, for most people, the performance metric ranges from -30% (35% on Windows) to +~ 30%. If you're doing better than +30% then you are exceedingly lucky and have an exceptionally rare workload which is near ideal for hyperthreading and your OS' scheduler. Which means, given an absolutely ideal workload, a six core hyperthreading CPU is roughly equal to eight cores. And if you're wrong, worst case, those six cores is roughly equal to a four core, non-hyperthreading system.

Re:AMD's stagnant?

[Belial6]

Every computer I have bought in the last 2 years has come with HDMI out, and except for the Acer Revo (nettop), they have all come with two video outputs. I understand that not everyone has HDMI on their TVs, but it is becoming less and less common to have those TVs. I would guess that the two reasons you mention are going to apply to a pretty small minority of people, as the kinds of people that will want TV out and multi-monitor support will likely have a computer newer than a couple of years old. If they do have an older computer, they generally will have already made that upgrade.

Re:AMD's stagnant?

[Antisyzygy]

The benefits of an integrated GPU on the processor die are not limited to graphics. You can tap the power of the GPU to compute certain things more efficiently.

Re:AMD's stagnant?

[Antisyzygy]

You must play older games.

Re:AMD's stagnant?

[godefroi]

15 years ago what you said there was true for both audio chips and network cards. Anybody who wanted one that was half-way decent bought a discrete unit because those performed well, and the hokey versions that you might find integrated were pretty much junk.

You're way way off-base here. The "integrated" audio chips and network "cards" you're talking about are still (and always were) discrete, they were just soldered onto the motherboard instead of stuck into a slot. The "integrated" graphics the GP is talking about are *ON THE SAME DIE*, as in, they *SHARE THE SAME SILICON* as the processor. Given the laws of thermodynamics, it seems unlikely that you'll ever have a high-TDP CPU and a high-TDP GPU (i.e. high-performance) sharing the same hunk of silicon. It doesn't make sense.

Re:AMD's stagnant?

[Belial6]

That depends on what you mean by older. I don't think I have any games made within the last year, although Plants vs. Zombies might have been, but many of them were made within the last few years.

Re:AMD's stagnant?

[cynyr]

who cares if it's last years Porsche, as long as it's fast, and competes on track days with this years, or on some tracks does better? Also what if you could get that 90% of the lap time of a Porsche GT for civic money? which would you buy?

From what I have read(I'm looking to buy a mini-itx system) on the issue of power is that both(1090T and 980x) will pull down around 130-180 watts at the PSU at full load, sure at idle the 980x wins, but if my computer is idle I probably should turn it off. I have a lower power consumption server for my "always on" tasks. If you want to compare power consumption, then ARM wins, for work per watt. My 8600GT will beat a i7-980x for some loads (highly parallel). Nvidia's Tegra2 stuff is pretty impressive performance/Watt.

On the HPC side, AMD has 3/10 spots, including #1 and #3. Intel as 4/10, #2, and #6. Granted the #2 Intel system is using ~1/2 of the CPU cores, and tesla GPUs, but Jaguar is still using DDR2-800 ram, and some PPC cores. Crays XT6m servers(#16) are using a 2.1ghz mangy-cours part and not the 2.2ghz part, gets 274.7 with 43660 cores, scaling up(no interconnect loss) to the same number of cores as the #2 machnine, i get 59.72% of the peek performance. I wonder how much that has to do with the tesla being better at linpack than the PPC compute unit on the XT6. Yes I'm aware that the interconnects and compute units matter a lot these days, but still #1 and #3 must count for something...

Cost is always a factor, even if you have near unlimited budget. As an mechanical engineer the question almost always is, "How can i get the same performance for less cost" or, "How can i get more performance for the same cost" Leaving cost out is like saying the Bugatti SS is a better car than a Subaru Forester. Now go camping with the family in Bugatti SS.

Re:AMD's stagnant?

[FreonTrip]

Having a GPU on the same die isn't technologically equivalent to the SIMD execution units of the CPU literally sharing resources with the geometry processors of the GPU, with which it has both figuratively and literally been, er, fused. I'm really quite intrigued to see how Fusion turns out, and can guarantee it'll be better than this water-muddying SoC Atom solution.

Re:AMD's stagnant?

[petermgreen]

The amount of stuff you can cram on a single chip is smaller than the amount of stuff you can cram on two chips, and chips that are twice as big are twice as likely to have catastrophic production flaws.
That is true, however interconnecting chips at high speed is also expensive and if you use a relatively low speed interface like PCIe then you end up needing to have seperate memory for the CPU and GPU.

Putting more cores in desktops is getting fairly pointless because most desktop apps aren't all that paralell and trying to make individual cores perform better by making them more complex seems to be hitting dead ends too. This suggests that at least on the lower end chips there should be transistor count and heat budget spare to add integrated graphics.

Re:AMD's stagnant?

[petermgreen]

You're way way off-base here. The "integrated" audio chips and network "cards" you're talking about are still (and always were) discrete, they were just soldered onto the motherboard instead of stuck into a slot.
Depends on the generation.

With audio Intel introduced a system called AC97 (later replaced by HDA which is similar in concept but supports higher data rates). There is still a dedicated chip (called the codec) to handle the final A/D and D/A conversion because analog stuff and high speed digital stuff don't mix well but any preproccessing and buffering of the data is either done by the southbridge or done in software.

A similar thing is happening with network though it only started recently and a lot of machines are still on the market that use PCIe or PCI network chips.

it seems unlikely that you'll ever have a high-TDP CPU and a high-TDP GPU (i.e. high-performance) sharing the same hunk of silicon. It doesn't make sense.
Notice how the CPU vendors essentially ran out of ways to make CPU cores faster a while back so they started shoving more than one of them on a die. If they can get 6 cores on a die now without sacrificing much performance per core over thier quad cores then I really don't see why four decent CPU cores and a reasonable (not top of the line but good enough to play most games on moderate settings) should be a problem from a power POV.

Re:AMD's stagnant?

[godefroi]

If they can get 6 cores on a die now without sacrificing much performance per core over thier quad cores then I really don't see why four decent CPU cores and a reasonable (not top of the line but good enough to play most games on moderate settings) should be a problem from a power POV.

Would you be willing to give up 4 of your 6 cores in order to have a "reasonable" GPU in the same silicon? That's what'd be necessary. Your own example shows that (desktop) CPUs continue to operate at the edge of the reasonable TDP envelope, and whenever there's room to spare, either more cores or more gHz show up to fill the gap.

Re:AMD's stagnant?

[bhtooefr]

I was trolling, I think you should've replied to the parent to my post. ;)

Will it be compatible with AM3?

[Liambp]

Will existing PCs with an AM3 socket be upgradeable to bulldozer, anyone know? I see a few people asking on forums and stuff but I couldn't find any authoritative answers.

Re:Will it be compatible with AM3?

[MrHanky]

No.

Re:Will it be compatible with AM3?

[WiglyWorm]

It will be drop in compatable with AMD server boards. At home, it will be AM3+.

Re:Will it be compatible with AM3?

[ponos]

Well, it seems it will be AM3r2 (AM+?). In keeping with tradition, bulldozer may be backwards compatible with AM3
as many people suggest in various forums (like AM2/AM2+). There is no evidence that it won't be compatible and
at this stage this is good news...

Re:Will it be compatible with AM3?

[hedwards]

AMD will keep it compatible, or at least release a version which is unless there's a technical reason keeping them from doing it. It's something they've had a history of doing and something which has worked out well for them in the past. Part of AMDs legacy is keeping things compatible wherever possible, for the release of their first AMD64 chips to the various addons they've included in recent years and the revisions of chips which could be used in previous boards. I strongly suspect that they'll keep the pattern up wherever possible, if for no other reason than it has worked so well for them in the past.

Bulldozer?

[Megahard]

Sounds like a slow-moving behemoth. Not the best choice for a name.

Re:Bulldozer?

[rotide]

Or it pushes everyone/everything else to the wayside. I guess it depends on your interpretation.

Re:Bulldozer?

[sideslash]

I think people's mental impression will vary; yours is typical of an office worker. To someone employed in construction or agriculture, a bulldozer is a symbol of getting huge amounts of work done in a very short time. This reminds me of an ad I saw some years back for an "object oriented database", where they showed a photoshopped race car with a tractor on the back end. Their marketing message was "Why do you have a sluggish RDMS on your web app's back end?" I found it hilarious, because my reflexive response was to ask, "Why is that totally useless racecar pasted on the front of that excellent looking tractor, the kind of vehicle that is used to grow all the crops that feed the world?" :)

Re:Bulldozer?

[omidaladini]

I wonder what it aims to bulldoze. Intel? Definitely not a good name.

Re:Bulldozer?

[Minwee]

Perhaps the team at AMD had been drinking heavily the night before.

At eight o'clock on Thursday morning Arthur didn't feel very good. He woke up blearily, got up, wandered blearily round his room, opened a window, saw a bulldozer, found his slippers, and stomped off to the bathroom to wash.
Toothpaste on the brush -- so. Scrub.
Shaving mirror -- pointing at the ceiling. He adjusted it. For a moment it reflected a second bulldozer through the bathroom window. Properly adjusted, it reflected Arthur Dent's bristles. He shaved them off, washed, dried, and stomped off to the kitchen to find something pleasant to put in his mouth.
Kettle, plug, fridge, milk, coffee. Yawn.
The word bulldozer wandered through his mind for a moment in search of something to connect with.
The bulldozer outside the kitchen window was quite a big one.
He stared at it.
"Yellow," he thought and stomped off back to his bedroom to get dressed.

If you don't know what that's from, get off the computer and go to a library.

Re:Bulldozer?

[RaymondKurzweil]

"Why is that totally useless racecar pasted on the front of that excellent looking tractor, the kind of vehicle that is used to grow all the crops that feed the world?" :

Maybe it's because the people that were selling that "object oriented database" were far more honest than you assumed.

Re:Bulldozer?

[sideslash]

It was thought-provoking, anyway. Nice when advertising can be like that.

Re:Bulldozer?

[maxwell demon]

So I get Intel will counter AMD's Bulldozer with their new Vogon architecture?

Re:Bulldozer?

[IndustrialComplex]

Sounds like a slow-moving behemoth. Not the best choice for a name.

If you've ever used a good bulldozer, you might be a slow moving behemoth, but the feeling you get is of an unstoppable juggernaut.

To paraphrase a popular parody commercial:

It get's shit done.

Re:Bulldozer?

[TeknoHog]

My impression of the name "Bulldozer" is something that gets a lot of shit done, but also spends a shitload of space and power. I thought both mobile and server worlds were moving into lighter and more efficient architectures, but I guess we just have to keep on using x86 for our closed software.

(Power-efficient x86 brings to mind names such as Intel Atom and Via Nano, but neither is particularly impressive when you compare them to real mobile/embedded architectures such as ARM and MIPS.)

Re:Bulldozer?

[Dutchy Wutchy]

This is what everyone should think about when they hear the name: http://www.youtube.com/watch?v=24n77GgRtrw

Almost certainly not AM3

[drinkypoo]

It's a whole new architecture and will almost certainly require a new socket. ISTR the article saying nothing about memory technologies as well. The good news is that a new architecture on the horizon which almost certainly requires a new socket makes it seem less likely that AMD will bring out another socket before then.

Re:Almost certainly not AM3

[JoeRandomHacker]

It sounds like the desktop version will be "AM3r2", or AM3 without the legacy DDR2 support.

Re:Sweeeeet nectar

[Pojut]

Cool, thanks!

Another article

[StuffMaster]

There is also this on the AMD site. It has a slightly different take on the core/module semantics.

Re:Another article

[bluefoxlucid]

I don't like the dual execution unit semantic. One ALU would actually provide optimal function with smart OOE microcode logic. Basically, one ALU, two inputs, and instruction classes. Obviously we can't MOV and MOV or MOV and MOVSB; but MOV and ADD might be plausible if they use entirely separate hardware. That may involve a fancy design for an ALU with some duplication; in fact, duplication of core components might make a good hybrid design (i.e. 50 instructions use the basic functionality of ADD, so we have two sets of the first few transistors in the pipeline reaching each branch, so we can ADD and SUB at the same time etc). Less overall die space == less power == less heat.

It hardly will change the processormarket..

[tramp]

because they can not really compete with Intel anymore. And Intel hardly punished for it's anti-competitive behaviour will laugh about it. Too bad but that will not change in the near future.

Hamster Farm Analogy

[lymond01]

Processor Speed: Very fast hamsters on well-oiled wheels
Multiple Cores: Many well-oiled wheels
On die memory controllers: dangled cheese
Cache: water trough next to the wheel
L3 Cache: Camelback packs for each hamster
Shared L3 Cache: This is where the real innovation comes in and won't be defined as patent is pending.

Re:Hamster Farm Analogy

[maxwell demon]

This is Slashdot. We need car analogies!

Re:Hamster Farm Analogy

[_avs_007]

Processor Speed: Very fast hamsters on well-oiled wheels Multiple Cores: Many well-oiled wheels On die memory controllers: dangled cheese Cache: water trough next to the wheel L3 Cache: Camelback packs for each hamster Shared L3 Cache: This is where the real innovation comes in and won't be defined as patent is pending.

Car Analogies:

Processor Speed: Max Engine RPM
Multiple Cores: Number of Cylinders in Engine (or number of gears in transmission)
On Die Memory Controller: Direct-Fuel Injection
Cache: Number of exhaust pipes before the CATs.
L3 Cache: Number of exhaust pipes after the mufflers.

Re:Hamster Farm Analogy

[toastar]

This is Slashdot. We need car analogies!

Well a train analogy would be a lot easier, the whole multiple engines sort of make since there, But I'll give the car analogy a go:


Processor Speed: Compression ratio
Multiple Cores: Number of Cylinders
On die memory controllers: 4 Wheel drive differential
Cache: Rim Size
L3 Cache: Tire Thickness
Shared L3 Cache: Flying Cars?

Re:Hamster Farm Analogy

[toastar]

Whoops

On die memory controllers: semi Auto-Transmission

Re:Hamster Farm Analogy

[DigiShaman]

Shared L3 Cache: Hamsters share the same cage and thus eat each others poop.

Re:Sweeeeet nectar

[ElectricTurtle]

Feeding the troll:

So, let's see, I can buy a 3.2 ghz hexacore from AMD for ~$300 or one from Intel for ~$900 on NewEgg right now... 5% price difference my ass. Even if it were 10% slower it would still be a killer deal.

Re:Nobody cares.

[Jesus_666]

Well, it's a new computer part. It computes. The computer I already own computes perfectly well so I don't see how a new one is any different. You act as if computers were a field of interest or even an industry.

What's next? People getting excited about new cars?

Oh yes you do, because the future is not desktop.

[Kupfernigk]

Power consumption. To use the discrete graphics card you need lots of bus drivers on the motherboard and the card. They have to drive the bus inductance and capacitance. The faster they go, the more bits on the bus, the more power they use. Integrate usable graphics and you are driving tiny loads, so the power consumption drops.

People simply don't want to sit in a fixed position governed by a box and a monitor, which is one reason laptops outsell desktops. The future is untethered, which means low power battery operated systems. Your discrete graphics card will never be more than a niche market.

Re:Sweeeeet nectar

[blair1q]

The Intel one does a lot more in 3.2 GHz than the AMD one does. That's the point of AMD needing a new architecture.

Pricing is not linear with performance, and never has been. If you have a performance advantage at the high end that your competitors can't approach, you get paid for it.

But if you want a ghetto computer, by all means buy one.

Re:Oh yes you do, because the future is not deskto

[0123456]

The future is untethered, which means low power battery operated systems. Your discrete graphics card will never be more than a niche market.

So no-one is going to play PC games anymore? I guess you could be right, but Microsoft better hope you're wrong.

Re:Same amount of cores = same performance?

[BlueFireIce]

Yes, because we all know you can compare CPU's because they both have the same amount of cores right? Wile the AMD Six-Core is allot cheaper than Intel's offering, that's because the AMD chip only competes with the quad core i7 chips (not with Intel's Six-Core chips, top AMD vs top Intel Six-Core chips, you would have to hit 4GHz+ OC on the AMD to even get close to the Intel at stock speeds) and for price reasons is far more accurate to compare with those. That is NOT to say that AMD's chips are not a a good buy, for the performance you get, the prices are quite nice, and depending on the use, might be better suited.

Via Nano

[naasking]

Honestly, I wish Via had the resources AMD and Intel have. Their Nano CPU is pretty nice, but it's languishing. They're only just now coming out with a dual core version. The Nano's on-die crypto extensions, low power use, and higher performance per watt would otherwise make it ideal for server applications, particularly SSL front-ends.

Doesn't follow

[Kupfernigk]

You have completely missed my point (you _have_ got electronic systems design experience, haven't you?). Given identical graphics capability, if you can put the GPU on the same die, or next door on the motherboard, as the main CPU, you will consume less power, as sure as V = Ldi/dt.

Anyone who has been paying attention for the last 10 years is well aware that the entire consumer electronics industry is largely driven by integration and shrinkage.

Re:Oh yes you do, because the future is not deskto

[boxwood]

You know laptops can still run PC games, right?

Re:Oh yes you do, because the future is not deskto

[imgod2u]

The next iteration of the XBox 360 will have an SoC that integrates both the GPU and CPU on a single die.

I think we're at a point where only a small niche (well, more so than before) pushes for the $600 behemoth video cards and $900 CPU's.

People are moving towards "just enough" machines that are light on price and power consumption.

Ghetto computer

[mark72005]

...which is all that the 99% of people who would be classified as "ghetto users" need anyhow.

Re:Oh yes you do, because the future is not deskto

[blair1q]

Even laptops are more comfortable to use on a desktop. They're actually annoying to use on a laptop.

Regardless, almost all the laptops you've ever used had separate CPU and GPU chips in them. So "untethered" is not driving integration.

What drives integration is price and the premium that can be charged for a lighter, smaller device. But the integrated chip will not have the same performance as separated chips at the same manufacturing cost. So you will get an integrated CPU/GPU based platform that costs the same as the old one but has less CPU or GPU performance.

Re:Nobody cares.

[hairyfeet]

Well I personally care because it affects my prices and thus the prices I charge customers. After the Intel douchebaggery came out, along with the Nvidia bumpgate fiasco I switched all new builds to AMD exclusively. The bang for the buck is just incredible, with OEM triples for $60 and quads for $99, and their IGPs frankly kick the dog snot out of Intel, with them even able to run games like Bioshock acceptably.

This new part is gonna be nice because it will allow for really nice SFF HTPCs, and with a Radeon onboard it will accelerate (I personally use Media Player Classic Home Cinema which works great with AMD GPUs) all the major video formats out there, thus taking the load off the CPU and allowing a nicer customer experience. I have found even the bottom of the line Athlon II duals thanks to the Radeon onboard give a really nice experience for the customer, and allows them to have all the Windows 7 Aero features without needing a discrete GPU. This will also be great for netbooks, although I haven't had any customer complaints from the new Amd Neo duals with Radeon onboard, this will lower power and thus lead to longer battery life.

So yeah, there are some of us nerd that actually care about such things. I like the fact that with AM3 being backward compatible I was able to replace my 7550 dual for a 925 quad without having to trash the box and start over, or that pretty much any of the 125w motherboards can have a 6 core dropped right in. This new Bulldozer will allow for PCs which use less power, generate less heat, while still giving a good user experience and great bang for the buck. While TFA says Bobcat is more for the low power I bet I end up using Bulldozer more, as folks like having that extra performance and the AMD "drag and drop" video transcoder makes it easy for customers to convert videos for their portable devices.

If you haven't checked out AMD in awhile you really should. Yes Intel has the absolute speed crown, but for everyday tasks and even heavy lifting like transcoding I've no complaints, and I certainly love having a fully loaded quad with 8Gb of RAM for under $650.

Re:Sweeeeet nectar

[Nadaka]

paying 1/3 as much for more than 1/3 the computing power is a viable strategy known as "value based judgment".

At any given price point where there exists an AMD processor, there are few if any intel CPU's with equal or better performance.

The i5 750 and i7 920 are among the very, very few intel chips that compete with AMD on value (performance / price).

Re:Sweeeeet nectar

[ElectricTurtle]

Benchmarks certainly testify that the AMD 3.2 ghz hexacore is not three times slower than the Intel. At three times the price even twice as fast is a rip off, and Intel isn't even that far ahead.

SIMD

[pgn674]

I worked with 128 bit SIMD (Single Instruction, Multiple Data) on an Intel x86 processor for my undergrad capstone, specifically SSE4.1. SIMD mainly allows vector operations. As one example, instead of adding 42 to a single 32 bit number in RAM, you can add 42 to four 32 bit numbers in RAM, if they're all next to each other, and do it in almost the same amount of time. Good for graphics and, well, vector operations. Kind of the CPU's answer to the GPU's specialties.

My capstone dealt with finding out if an ignorant person who otherwise knows assembler can use SIMD in general purpose cases to speed up his program. The answer is 'probably not', as SIMD tends to mess up branch prediction, pipelining, and out of order execution, and it can take significant overhead to make SIMD work for you. Best to stick with what it was specifically built for if you're going to be using SIMD.

Re:Sweeeeet nectar

[blair1q]

Paying 3X as much for 2X the capability and profiting 4X as much from it is a viable strategy known as "making more money".

Seriously, if all you want is enough computer to post to /. on, you have no business worrying about CPUs more recent than the Celeron and Duron.

Re:Sweeeeet nectar

[blair1q]

By that logic, you should be buying old Pentium-D chips for $5 a pound and bragging that you've got more performance/$ than the new Bulldozer will have.

Plus, Windows Server support is over...

[BUL2294]

Even Microsoft has thrown in the towel with IA-64 given the scalability of AMD64 (err, x86-64) in Xeon & Opteron processors. Windows Server 2008 R2 is the last version to support IA-64...

http://www.microsoft.com/windowsserver2008/en/us/2008-IA.aspx

Re:Oh yes you do, because the future is not deskto

[hedwards]

I misread that at first, yes you're correct, discrete cards are somewhat problematic in that respect. However, it's not quite that bleak, with technology that's been coming for a while, it wouldn't surprise me if before too long the integrated on die graphics chip was teamed up with either an integrated or more likely discrete graphics card to somewhat bump the quality even further.

But there's a reason why AMD wants to do it, they've had a lot of good luck with integrating things and strategically moving and removing unnecessary buses. Not completely unlike what Woz was doing back in the late 70s by removing and consolidating chips in the original Apple computers.

Re:Sweeeeet nectar

[Runefox]

There does come a point where the increase in speed is not at all worth the extra cost. The top-tier AMD parts (the hexa-cores and the 965's) are more than capable of handling any task effectively, and they represent about the mid-high range of the Intel line in terms of performance (high i5/low i7) for about the same price - the 965 Black Edition is more or less on par with the i5-750, and is $30 less expensive. Calling AMD "ghetto" as you have in other posts is wholly incorrect, and remember that it wasn't very long ago that Pentium 4's were being smoked by Athlon64's selling for half the price.

But it comes down to a matter of usage. The i7's are ridiculously fast processors, but most people won't ever need that. The fastest i7 isn't even 50% faster than the fastest AMD part, and it runs a good 250% or so more at $1,176.14 (vs $334.99) on NCIX. For someone like me, I'd love a high-tier i7, but to be totally honest, there just isn't any compelling reason for most people to crack open their wallets for that. Intel blows away the high-end market with the i7's because nothing can touch them, but AMD has the midrange-low end market cornered, particularly with the rather decent Athlon II series and affordable Phenom II's.

Re:Oh yes you do, because the future is not deskto

[Z80a]

hm, i think the "next iteration" you're talking about was launched already, its that 360 slim/s.

Re:Sweeeeet nectar

[ElectricTurtle]

And by your logic, it would make sense to buy something for three times the price just to get a quarter more points on your PCMark score, just because that's the fastest possible. That's the way insecure men have mid-life crises. "I don't care what it costs, so long as people think I have the biggest penis!" Yeah, that's Intel consumers alright.

Yes, please.

[KonoWatakushi]

Just about every iteration of SIMD from Intel and AMD has been utterly worthless. (Not to mention NEON on the ARM.) Altivec was an example of SIMD done right, and AVX finally incorporates some of the better features of it.

Re:Yes, please.

[Bengie]

yeah.. SIMD sux.. My video compression is only 3xs faster with SSE... booo!!!

And to some people

[Sycraft-fu]

Performance per chip, or per core, is important. More cores is nice and all... If you can use them. Not everything can. If your apps use only 2 cores, the other 4 don't do you a lot of good. However maybe those apps need a lot of performance out of the cores they do use (games are like that). As such you are interested in performance per core, not just having more cores.

Re:And to some people

[blair1q]

If you're running any Windows since XP, you're using all of your cores all the time, and it's benefitting you. You may not get all of them working on the same task, but the fact that the other cores can do background and response tasks while your foreground task pegs one of the cores is always a bonus. If it doesn't show up in outright speed of completion, it hides a vast array of niggling little delays that make things jerky.

Re:And to some people

[macshit]

OTOH, if you can take advantage of many cores (I can), AMD's low, low, prices for lots of pretty fast cores are wonderful!

Re:And to some people

[godefroi]

There are surprisingly few (on a well-administered system) "background and response tasks" that consume a significant amount of CPU time. In general, pretty much everything is waiting on I/O of some sort, and having more cores doesn't reduce that wait.

Re:And to some people

[plonk420]

me too, but with case, psu, ram, hdd, i7-920 (and moreso 860) was a better deal for about a year and a half. there are SOME distributed computing (that's not ALL i do, fyi) people that run their farms on paper trays or other neat little vertical stacks, though.

Re:Nobody cares.

[Profane MuthaFucka]

I've read the Bible. God "blessing" anybody is the biggest fuck you of all.

Re:Sweeeeet nectar

[Lehk228]

$1176? my entire 8gig phenom 965 black edition machine cost that much including the SSD primary hard drive

Also they got psychological advantage

[Ilgaz]

i7 may be damn good with way more higher technology and support but really, its price is dumb. Really dumb. One can't really justify paying that price (don't forget decent MB) for a desktop processor. Lets not forget the integrated ATI graphics , even lowest end can show Intel GMA as a joke. At least they got hardware T&L in 2010 for God's sake.

I have serious problems paying $1000 for a processor and I own a Quad G5 PowerPC from Apple. Now if it makes me feel like that, imagine the rest. Also one day, one non dumb company will manage to make users really use "cloud" computing all with grid technology with privacy concerns kinda resolved, that time netbook owners may really laugh at $1000 CPU owners for a reason.

If Apple didn't produce hardware and didn't make living with hardware upgrades, they could be the ones to solve this stupid problem with an elite and yet practical solution. In fact, every OS X comes with it enabled... XGrid...

Re:Oh yes you do, because the future is not deskto

[Nutria]

People simply don't want to sit in a fixed position governed by a box and a monitor

Get Off My Lawn!!

19" monitors, full-sized keyboards and real optical mice are much more comfortable than laptop h/w.

Besides, I'm disabled and can only use one hand to type and thus laptop keyboards suck.

Makes me wish that Byte magazine still existed.

[Dante]

Reading this makes me so nostalgic for byte magazine, they on occasion would right the most awesome, and in depth articles about processors. They did not just regurgitate the companies white paper, and glossy marketing copy.

Re:Nobody cares.

[Alcoholist]

You're on glue mate. Selling AMD is a ghetto.

Don't misunderstand, I used to build and sell OEM AMD based systems religiously for years and what did it get me? I've had more dead boards and more dead CPUs than I ever had with Intel stuff. And it's not like I was using crap mobos either, I was using Asus and MSI. I was using Kingston RAM, Antec PSUs and these things were still falling over left and right. I've lost so much money servicing the warranty on their platform it's embarrassing. Intel, out of the hundreds I've assembled, I've seen less than a dozen back. They just work, no complaints, no problems.

I'm not sure what 'customer experience' is supposed to mean, but me not going out of business matters to me somewhat more. There is a reason why they sell their CPUs for $99 these days: because they are slow and shitty and no one with any brains wants to buy them.

cross-licensing and antitrust laws

[G3ckoG33k]

I think the cross-licensing of x86 and amd64 technology and antitrust laws have something to do with it, too.

Re:Nobody cares.

[hairyfeet]

I'm sorry, but if you are actually managing to blow ANY MODERN CPU then "Ur doin it wrong".Both CPU and board have excellent thermal management built in, and short of the gamer boards they are usually set by default to be VERY conservative. The only OEM AMDs I've ever seen blown were where the builder himself went "ghetto" and didn't even bother to stick in a single fan other than the CPU heatsink, which they went bottom of the barrel.

The ONLY difference I've found between modern AMD and Intel when it comes to builds is you can really cheap out on the lower end Celeron as far as fans are concerned and they won't overheat, thanks to Intel learning from Netburst and overdoing it when it comes to heatsinks. So if you are seriously blowing Asus and MSI boards, which are both good brands, then I'd look to see what mistake you are making, most likely you are either OCing too high (I have seen this quite often, where an OEM will OC a PC to sell it as a faster chip than it is) or are skimping on cooling. Because I have built both Intel and AMD PCs for some truly rough environments, like uncooled warehouses, construction trailers, lumber mills, etc, and never had a single build blown by the CPU. Had a few bit by the bad caps that went around, but that wasn't Intel nor AMDs fault.

So if you want to charge your customers more money for power they'll frankly probably never use, or just prefer building one over the other feel free. But I haven't seen either company make board frying chips since the early Netburst days, so if you are seriously blowing boards and not just trolling you might want to look again at the BIOS and see what you are setting wrong or maybe add an extra fan. It isn't like an extra 80MM will set you back a fortune, and I've found customers prefer a PC that stays nice and cool even when doing heavy lifting. But if AMD were putting out board blowing CPUs it would have been all over the news by now, and as many as I've cranked off in the past couple of years I'd be seeing returns which I honestly haven't. In fact I've not had a single Intel or AMD returned to me for warranty work in more years than I can count, but I don't skimp on the fans or OC either. The only time my boxes come back is when they want an upgrade, which is how I and my customers like it.

Re:Nobody cares.

[dbIII]

I'm sorry, but if you are actually managing to blow ANY MODERN CPU then "Ur doin it wrong".

Funny this this morning is I had a user complain about their computer running for about 5 seconds on start then shutting down. I opened up the case and the heatsink fell out, the clip mounting point had snapped off.
The thing was running with nothing but a bit of paste on top and shut itself off safely.

Re:Nobody cares.

[GSV Eat Me Reality]

  All due respect, but my experience is exactly the opposite of yours.

  I've custom built and sold over 800 systems in the last five years, and the AMD based systems, in terms of hardware reliability, are far, far better than the intel based systems - especially in the mini case formats, where heat is a factor, but including the gaming rigs.

  Just looking at my invoices for this year, hardware replacement for the intel systems is over three times as common as it is for the amd based systems - and that includes the work I've done on systems I didn't build.

  I'm not even going to go into mobo driver issues, except to say fuck ATI. Any cognitive diss wrt that you can pipe to devnull.

  I agree with the GP. AMD kicks ass. I have seven custom built systems at home, all AMD/nvidia combos now as of early this year, because Intel's kit has not treated me very well since at least '05. They are rock solid. Downtime? What's that?

  YMMV, Flame on.

  GSVEMR

 

ARM worshipping is annoying

[renoX]

Do you realize that ARM doesn't even have a 64 bit version of the ISA??
MIPS and PowerPC have them.

Also AFAIK it doesn't have a 'trap on integer overflow' mode like MIPS:
very interesting for a language like Ada which has exception on integer overflow:
on MIPS this feature could have zero performance impact during normal execution.

Re:ARM worshipping is annoying

[dkf]

Do you realize that ARM doesn't even have a 64 bit version of the ISA?

Since they're strongly targeted at the embedded market, why would the ability to deal with more than 4GB of memory be any use? (In time it may come to be useful, but within the next, say, 3 years?) ARM processors own their niche, and it's a gigantic niche with good long term prospects. Why worry about competing with Intel in a different area?

Re:ARM worshipping is annoying

[renoX]

>why would the ability to deal with more than 4GB of memory be any use?

Yeah, no use at all to ARM of course, hhmmmm, maybe you should read this: http://www.theregister.co.uk/2010/08/25/arm_server_extensions/

Quite a few people have expressed the desire to have servers or laptop with ARM CPU, in both case more than 4GB memory is useful.
Hence the ISA extension of ARM to 'PAE like' 40 bit, of course this is a kludge as each process cannot access more than 4GB..

Re:ARM worshipping is annoying

[imgod2u]

ARM just announced a 40-bit extended addressing mode and their roadmap is for a 64-bit extension in 2012. I'm not sure how useful traps are in this day and age.

Re:Nobody cares.

[dave420]

They used to kick ass, but recently AMD's offerings, while cheap, have been totally shit compared to Intel's offerings.

Re:Nobody cares.

[hairyfeet]

Exactly, which was why I was pointing out the guy supposedly blowing out CPUs was "doin it wrong" as modern boards and CPUs are VERY conservative when it comes to temps. The last time I can remember actually seeing a blown CPU was the (I believe Barton) very old AMD chips that had NO thermal management, and the first P4s that would run a hell of a lot hotter than the P3s. Even then it was the fault of whomever built the box, as they were skimping out on fans and using junk quality heatsinks. I myself have NEVER, in all my years building PCs, going back to Win3.x, managed to blow a CPU.

The trick I've found in really building rock solid reliable PCs is to avoid the "gamer kit" boards, as many of them are set with higher voltages from the factory to give it a "natural OC" which gets better scores on benchmarks. I start with an ECS, MSI, or Asus Business Class board, as those are designed for long life with much better caps than consumer class, then I do NOT OC, use a quality heatsink and at LEAST two fans (if in a rough environment I got negative flow, otherwise through and through) and then use quality parts like Crucial or Kingston RAM and WD drives. It may cost a little more than building a "ghetto" box, but my customers appreciate that they can pass one of my builds onto relatives when it comes time to upgrade and it keeps on ticking.

As a matter of fact speaking of AMD I had a 1.2Ghz Duron I built for a local minister ages ago come in right before vacation. I thought for sure the Rev would be ready for a new box, but since I bumped him up to 1Gb in 06 he said he couldn't be happier and only wanted a bigger HDD added so he could have room for the audio/video he has been taking of the choir. This "ghetto AMD" as he would have called it has been running pretty much 24/7 since 2001 without a hiccup. I also have a 1.1Ghz Celeron I built at the same time currently being used by a local checkout girl, after it passed through half a dozen relatives in her family. Other than being upgraded to XP in 05 it is the same as it left my shop, no problems.

So if I had to guess I'd say Mr "AMD is ghetto" was/is doing a slightly dirty trick I have seen other builders do I call a "factory fuck u" where they OC the chip, while skimping out on even a case fan, and then sell the PC as a faster chip than it really is. I recently worked on a factory FU where the guy thought he had bought a 1.5Ghz AMD in 06, when actually it was a 900Mhz the asshole builder had cranked an OC on and sold as a 1.5Ghz. Needless to say as the fans wore down it became more and more unstable, eventually trashing his data. It really is just shameful that some will resort to such underhanded tricks to make a little extra profit, but maybe I'm just weird with my "cost plus 20%" rule. I won't ever be rich but I take pride in the fact my boxes last a looooong time.

Re:Sweeeeet nectar

[AmiMoJo]

Since when has there been a linear price/performance relationship? It's always been a curve, that's the way it works with technology. If you want to go 10% faster you pay more than 10% extra, if you want a 10% thinner phone you pay more than 10% extra etc.

Re:Sweeeeet nectar

[ElectricTurtle]

I'm not saying it's linear, I'm saying that at a certain point the increase isn't worth it for most people. That there is a curve is real an undeniable, that things at the top are worth their cost to any but a few special cases (and the insecure) is unlikely.

Re:Oh yes you do, because the future is not deskto

[godefroi]

I just recently pushed for a "behemoth" video card, but it was only $150. Am I a small niche?

Re:Sweeeeet nectar

[AmiMoJo]

That was my point. You can't look at it as 2x the cost for 2x the performance improvement, you have to look at is as wanting cutting edge technology. Calling it a rip-off is not really fair because it isn't a deliberate over-pricing, it's due to lower manufacturing yields and higher manufacturing costs.

Re:Sweeeeet nectar

[ElectricTurtle]

Calling it a rip-off is not really fair because it isn't a deliberate over-pricing, it's due to lower manufacturing yields and higher manufacturing costs.

If you truly believe that, I have a bridge to sell you. Continuing to use the example of the AMD 1090T, it is AMD's flagship desktop processor, only on the market since May, and even when it was in shortest supply the price was still only $400 maximum [citation]. The Intel 980X by comparison has been out longer, everybody knows that Intel has a larger manufacturing base, and still their price hasn't dropped by as much in terms of percentage[citation].

Intel is leveraging its brand identity to keep their prices inflated, even though their ability to supply is known to be superior. The people who pay that cost are a) not spending their own money, making decisions for a business or government or b) insecure and trying to keep up with the Joneses or rarely c) have a special application which absolutely requires a minor increase in speed at three times the cost.

Re:Sweeeeet nectar

[AmiMoJo]

I wasn't arguing that Intel as a brand isn't overpriced, although even that is not always the case. If you compare branded budget laptops and desktops Intel based ones are actually fairly competitive, not least because there are a wider variety of chipsets and motherboards that support the Intel platform so motherboard prices can even out the total cost.

What I was simply trying to say is that the pricing curve is not, as you suggested, completely artificial. Nothing more. Sure Intel charges more than AMD for a similar product, but what does that have to do with high end parts costing exponentially more to produce? You can argue which CPU is better all you want, I never disputed that.

Honestly, it would save a lot of time if you just bothered to read the bit you actually quoted in your reply. Perhaps then your reply might have actually addressed it rather than going off on a tangent about what you thought my point was in your own mind for some inexplicable reason. FWIW my current CPUs are Intel (Core 2 Quad, Pentium Dual Core and an Atom 510) but only because they were cheap, and my next will probably be an AMD because total cost (CPU+mobo) is lower and AMD's chipsets are robust with decent on-board video.

Re:Nobody cares.

[toddestan]

Are you sure it's not just the components? I know Antec and Asus are well regarded, but I've never found either brand to be reliable, both on AMD and Intel systems. Certainly not worth the price premium they charge.

I'll still give Intel the nod on reliability though. Intel boards with Intel chipsets are absolutely rock solid from my experience. This is coming from someone who's last 3 builds are AMD.

Re:Nobody cares.

[badkarmadayaccount]

Whoosh!

Re:Sweeeeet nectar

[badkarmadayaccount]

How dou you think clusters work (though they don't usually use P-D chips.

Re:Nobody cares.

[hairyfeet]

I'd say the same thing of AMD boards with AMD chipsets. I haven't seen a single driver problem with ATI chipsets since the 8 series, and since AMD took over they have really worked hand in glove to make a pretty damned rock solid product. My personal rig hasn't even been shut down except for a passing storm or two since Oct of last year, and the Windows reliability monitor shows NO errors from drivers.

As for boards I have to agree about Asus, I've found most of them are designed to run "hot" or be slightly overvolted from the factory, which lets them score higher benchmarks but really lowers reliability and life. I'll probably get laughed for saying this, but I have had nothing but good luck with the ECS Business Class boards. Not only are they better priced than the Asus boards, but they use solid caps and have all around rock solid designs. The only thing is they are business boards, not gamer rigs, so you only get one PCIe x16, but frankly I think Crossfire/SLI is overrated and isn't worth the extra heat and power, and most of my customers couldn't care less as the onboard does all they need. I just dropped a 4650 1Gb in mine and it plays Wolfenstein, Bioshock 1&2, and every other game I've thrown at it just fine.

So next time you build an AMD you might want to look at the ECS Business Class. As long as you aren't wanting Crossfire they are really solid, with plenty of options in the BIOS, most support from 16-32Gb of RAM, just really nice solid boards. I've been using them for a couple of years now and if anyone could kill them my customers could, so I'd say they are pretty solid.