AMD QuadFX Platform and FX-70 Series Launched

MojoKid writes, "AMD officially launched their QuadFX platform and FX-70 series processors today, previously known as 4x4. FX-70 series processors will be sold in matched pairs at speeds of 2.6, 2.8, and 3GHz. These chips are currently supported by NVIDIA nForce 680a chipset-based, dual-socket motherboards, namely the Asus L1N64-SLI WS, which is currently the only model available. HotHardware took a fully configured AMD QuadFX system out for a spin and though performance was impressive, the fastest 3GHz quad-core FX-74 configuration couldn't catch Intel's Core 2 Extreme QX6700 quad-core chip in any of the benchmarks. The platform does show promise for the future, however, especially with AMD's Torenzza open socket initiative." And mikemuch writes that the QuadFX "not only fails to take the performance crown from Intel's quad-core Core 2 Extreme QX6700, but in the process burns almost twice as much electricity and runs significantly hotter in the process. ExtremeTech has a plethora of application and synthetic benchmarks on QuadFX, including gaming and media-encoding tests."

Comments

Also in the AMD vendor section

[Anonymous Coward]

AMD unveils processors for 'power' computer users

Boy, intel really made things difficult.

[3.5 stripes]

From core 2 duo to core 2 extreme, just doubling the cores.. how are they going to top that with 8 cores? Core 2 tubular?

Re:Boy, intel really made things difficult.

[oliverthered]

Not 8 cores 80 cores

Re:Boy, intel really made things difficult.

[Anonymous Coward]

The dyadic core2 double deux duo, naturally...

In other words a redundant repetitious repetitive name that seems to repeat it's self.

Not long now

[LiquidCoooled]

Until we replace our kitchen hobs and kettles with computer CPUs.
Having the home server also doing all water heating for the house might be a good idea.

You need to run some intensive process to heat up enough for a bath or shower.

Re:Not long now

[WhoBeDaPlaya]

Folding or Prime95. 'Nuff said ;)

Re:Not long now

[Anonymous Coward]

Coming soon to YouTube : How to use Quad FX for Korean BBQ

Re:Not long now

[Anonymous Coward]

Folding or Prime95. 'Nuff said ;)

That's the hilarious thing about all these hardware stories. Like you guys have any possible use for this additional power besides running pointless benchmarks.

Re:Not long now

[Dunbal]

You need to run some intensive process to heat up enough for a bath or shower.

      Nahh, just boot Vista.

Re:Not long now

[chrismcdirty]

I don't think many people would call Folding pointless.

Re:Not long now

[0xABADC0DA]

Folding a point is an identity operation so really it's more futile than pointless.

Re:Not long now

[Mikachu]

You wouldn't call Vista intensive process? ;)

Re:do they read what they write?

[Anonymous Coward]

>>How about reading the article after you write it?

They're in the process of reading it now.

Good name from a marketing perspective

[traindirector]

While performance may be disappointing, it's pretty clear that AMD is just releasing this as a stopgap solution to "stay in the game" for the performance sector until their new developments are ready next year. The name is a good choice and reflects that intention - they combine their performance branding, FX, with "Quad", the term Intel is using, to indicate that it fills the same niche as a quad-core processor. I think it does what it is meant to do - give the impression of a comparable offering until AMD has the real competition ready.

Re:Good name from a marketing perspective

[tomstdenis]

Lame though...

I'd rather an Opteron for HPC than an Intel box anyday.

For "gaming" and other 1337 chores, I guess the Intel box is the winner. But really until Intel figures out this "not use the FSB" approach they can kiss my HPC using ass good bye.

[and this is coming from a dude who loves his Core 2 Duo workstation...]

Tom

Re:Good name from a marketing perspective

[Anonymous Coward]

Sooo... when Intel had the hotter, more energy wasting, slower processor... it was "omg look how much Intel sux0rz" but now when AMD is in that boat its... "this is just a stopgap, you just wait!"

Fanboys... it just doesn't get more entertaining then that.

Re:Good name from a marketing perspective

[afidel]

Well, it depends on what your doing for HPC, if you have something with a high FP workload to messaging/data workload then the Intel part might be a good choice since it is by far the Floating Point king. If your workload consists of large data sets or lots of message passing then the AMD solution might be the right fit. I deal with more classical IT workloads (large database and n-tier systems) so the AMD solution is currently the better fit for my heavy lifting boxes. My blade servers for things like Citrix servers use dual core Intel chips because of the currently better MIPS/Watt ratio.

Re:Good name from a marketing perspective

[traindirector]

Sooo... when Intel had the hotter, more energy wasting, slower processor... it was "omg look how much Intel sux0rz" but now when AMD is in that boat its... "this is just a stopgap, you just wait!"

Intel spent years in that boat with no indication that they had an intention to 1) aim for low power consumption (they were happily gloating about the forthcoming Pentium 4 5GHz) or 2) do what it took to gain the performance crown. It was not clear (in recent history) that they had an eye on the super-performance desktop market until the announcement of the Extreme Edition and little indication of concern about power usage on the desktop until they announced that their new desktop processors would be based on the Pentium M.

On the other hand, we already know AMD's plans for next year, and we have statements of what they hope to achieve. I'm not saying just to wait and that it will be awesome. I'm posting on a Core 2 Duo system built using the remnants of my last Athlon XP system. My previous post indicated my expectations for what AMD is doing from a business perspective, not my feelings about the company or their product.

Re:Good name from a marketing perspective

[Lonewolf666]

It is rather obvious that the QuadFX is a niche release for gamers, on several grounds.

1) Only one memory module per channel supported, and no ECC. Not what you want for high reliability or in memory-intensive tasks. This disqualifies the QuadFX as server processor.

2) Much cheaper than the (closest in performance) Opteron 2220 SE but castrated as described above. I guess AMD did that intentionally to avoid cannibalizing the server market.

So I'd consider the Intel, with a high quality board, for "affordable HPC". Not the QuadFX. Disregarding bigger machines for a moment. If you have the money, an 8-way Opteron looks pretty nice...

Re:Good name from a marketing perspective

[RightSaidFred99]

I can only assume that by "give the impression" you actually meant "give the illusion".

I'm actually surprised the review was so positive. It's somewhat neutral, but really the conclusion is simple - 4x4 is a piece of crap. It uses ridiculou power and still loses the majority of benchmarks. In fact, in most applications, an E6600 would match or exceed it in performance, especially when both are overclocked.

Espensive, high power requirements, average performance, questionable future. Seriously, pretty sad offering.

Re:Good name from a marketing perspective

[saider]

Don't forget...

Piss off gamers with a problematic part and you might lose some "street cred".

Piss off IT managers with a problematic part and you will lose significant revenue for many quarters to come.

If I were going to test out a new product, a bunch of rich kid early adopters would be the market segment to target. They are always willing to try something new and their decisions do not significantly impact your bottom line.

Once the process kinks are worked out, incorporate the other features for your main line processors.

Re:Good name from a marketing perspective

[pseudorand]

It may be true that it's just a "Stay in the game" measure, but that won't stop me from buying Intel from now on. I've had horrible heat problems with AMD's 280 and 285 dual-core CPUs and 854 single-core CPUs (at least on Tyan boards, the only ones I've tried), and now they're asking me to pick up a quad core that's slower, hotter, and has to run on a linux-hating (try temp monitoring) Asus board? I think not. This is enough to make me switch to Intel. And when AMD's ahead in the benchmarks, I'm not going to make the mistake of switching back again. Once bitten, twice shy.

Re:Good name from a marketing perspective

[Anonymous Coward]

Piss off IT managers with a problematic part and you will lose significant revenue for many quarters to come.

As an IT manager one thing thats glaring at me is this thing uses two sockets, the same measure many venors use to license expensive code. So an Intel 8-core box would cost half of what a AMD 8 core box would to license. Which means AMD loses market share anyway. (Not to mention this thing doesn't seem to support 8-core operation yet anyway :(

Re:Good name from a marketing perspective

[Lonewolf666]

Not to mention this thing doesn't seem to support 8-core operation yet anyway :(
Correct. It appears to be a higher clocked but otherwise severely limited version of the Opteron 22XX series, relabeled and fitted with a controller for non-registered, non-ECC RAM.
The QuadFX may be interesting for gamers, but I doubt anyone else would want it.

AMD's new Power HOG

[Gr8Apes]

yep, it's a power hog. 400W @ idle? Youch!

But, this is a first release, and what's important is the strengths shown. Notably, that 2 AMD 64 processors (granted, the 1207 pin versions) scale up to Intel's brand new Core 2 QX series best (itself 2 CPUs slapped together). It will be interesting when AMD releases their true quad core CPUs on 65nm in 2007. It looks like they'll be on par with Intel at worst.

This is only good news for us consumers!

Re:AMD's new Power HOG

[edflyerssn007]

This is true, while Amd lags now, it is still on 90nm while intel is on 65, when both are on 65 then we'll have some real competition going on. -Ed

Re:AMD's new Power HOG

[MonaLisa]

Only by the time AMD releases their 65nm stuff, Intel will be at 45nm http://today.reuters.com/news/articleinvesting.asp x?view=CN&storyID=2006-11-27T225533Z_01_N27466640_ RTRIDST_0_INTEL-MANUFACTURING.XML&rpc=66&type=qcna

Re:AMD's new Power HOG

[fitten]

Notably, that 2 AMD 64 processors (granted, the 1207 pin versions) scale up to Intel's brand new Core 2 QX series best (itself 2 CPUs slapped together). It will be interesting when AMD releases their true quad core CPUs on 65nm in 2007.

"true" again... spoken like a "true" fanboi. Anyway... Core has a lot of headroom for frequency increases. In the ExtremeTech article, they overclocked the QX part from 2.66GHz to 3.55GHz even. Not only that, but Intel is already sampling 45nm parts and will likely have 45nm Core parts out by the time AMD can ship enough 65nm to really matter to the market. Along with 45nm comes even more clockspeed increases (rumors are of, or nearly, 4GHz rated Core parts, not overclocked, sampling now).

*WE* need AMD to catch up so at least we have competition in the market again. I buy what's best on the market. For the past few years, that was AMD and I have 6 AMD machines (four AMD64/X2) because AMD was the best. It doesn't look like I'll be buying any more AMD parts for a while, even with the 'release' of this 4x4 junk. Barcelona remains to be seen, but the day (or the deay before) it's released, I'm sure Intel will release a few frequency bumps that will take a lot of the shine off the new AMD parts.

Re:AMD's new Power HOG

[Svartalf]

Ahh... Someone that gets this concept.

Intel wins on extra Cache- and the benchmarks that keep getting ran don't reveal performance snags with the SMP operation.

Intel's got a shared L2 that's 2-4 times the size of the AMD equivalents' pool.
AMD's got a coherent, but NON-shared L2 split across multiple CPUs- each core has it's own L2. You'll have less L2 thrash with that design.

Under an SMP load, the AMD design will have an edge if all four cores are busy in different parts of system memory.
If you pop out of cache, the memory bus design and overall architecture of the AMD parts will have an edge.

Intel has an edge only due to process shrink and the things they can do as a result thereof. As soon as AMD goes to the
smaller process size, they'll pick up the lower TDP advantage Intel has right at the moment and then the whole deal will
flip-flop on who's got the "best" CPU unless Intel comes up with a few new tricks along the way, which may/may not happen
for them.

Re:AMD's new Power HOG

[fitten]

overclocked the QX part from 2.66GHz to 3.55GHz even.

My bad... I read that wrong. The Core2Duo was overclocked to 3.55GHz, not the quad core part. However, this still shows the amount of headroom the core has on 65nm.

Re:AMD's new Power HOG

[fitten]

Intel's got a shared L2 that's 2-4 times the size of the AMD equivalents' pool.
AMD's got a coherent, but NON-shared L2 split across multiple CPUs- each core has it's own L2. You'll have less L2 thrash with that design.

Under an SMP load, the AMD design will have an edge if all four cores are busy in different parts of system memory.
If you pop out of cache, the memory bus design and overall architecture of the AMD parts will have an edge.

In CPU architecture circles, the shared L2 is considered a more ideal design than split L2 for multi-core processors. There are plenty of talks around the 'net as to why.

As far as cache size, that's a design tradeoff just like any other. Because of the slowness of main memory, you want to have as large a cache system as possible. However, cache system latency increases with the size of the cache so that is a tradeoff as well. Intel chose to use some chip realestate for cache. "Faulting" them for this is just being an apologist for your puppy.

There are many types of "SMP loads". Multi-threaded loads where all threads work on the same data will be similar on both as there is only one pipe to the memory on both the NUMA and the FSB model, for example. But yet, on SMP loads that are more 'lose', you can get good benefits from NUMA. By the way, Intel also has the IMC with their equivalent to HT on the roadmaps, so this discussion (NUMA vs FSB) won't be relevant for much longer.

Additionally, it isn't until AMD's 'next thing' where their NUMA architecture will be able to scale much better (it doesn't do that well with lots of sockets because it falls back to being limited by the number of HT connections so some communication has to be multi-jump with current multi-socket solutions - the new core adds an HT link so that 4+ sockets can have a more direct path around the system).

There are a number of examples of "popping out of cache" in the tests on various sites. AMD does show that it helps in those when it can use the bandwidth of both NUMA branches but it isn't convincingly better than Intel's FSB on many/any of the tests that are shown (you'd hope to see idealistically 2x performance improvement on many of those, but even with all the extra bandwidth, AMD doesn't seem to 'blow the doors' off of the Intel parts... in fact, AMD doesn't even beat them even with the added bandwidth... this just shows that there may be more to the picture than an IMC + more bandwidth). Even AMD's latency isn't that much better than Intel's FSB design anymore (the nice advantage that had against NetBurst is pretty much gone).

I'm eagerly awaiting AMD's next 'real' move, myself, but given that Intel is already sampling 45nm parts and even on 65nm Core is able to clock to 3.5GHz ranges (meaning Intel has a lot of headroom even on 65nm), the short amount of time that Intel and AMD will overlap on 65nm will probably just show equality (at best) between the two. I haven't really seen what performance advantages AMD's new features give, other than the obvious benefits of wider paths and the FPU issue increase (to bring it equal to Intel's issue rate, although AMD has typically had a stronger FPU). AMD claims a lot, but that could simply be marketing at this point.

Re:AMD's new Power HOG

[geekoid]

Throwing a crap product on the market as a 'stopgap' is not good for consumers.

OTOH, this is typical AMD. Copy Intel, but hotter and a power hog, then try to refine. About the time AMD is actually a real competitot to the Intel version, Intel cmoes out with a better chip and AMD follows. Rinse repeat.

Re:AMD's new Power HOG

[geekoid]

Then Intel will come out with something else, and AMD will be playing catch up again.

Re:AMD's new Power HOG

[sxpert]

the announced quad core AMD has a shared L3 on the chip itself...

Re:AMD's new Power HOG

[Gr8Apes]

1) I'd buy a Core 2 today if I needed a performance part. No "fanboi" here. Note my comment was about competition and what's good for consumers.

2) The increases seen from dropping mask scale haven't been that great over the past few drops. Certainly not in the frequency bump arena. Since 65 can do 3.6 or more, I wouldn't be shocked if they can eek 4GHz out of 45nm, finally, only about 4 years late.

3) Intel's architecture has already shown problems with the FSB limitations with Woodcrest in Apple benchmarks. True SMP type multi-threading causes degraded performance, especially when heavy large-scale memory access occurs that exceeds the cache capacity. I can't imagine this will get better @ 45nm, except that the cache might grow. 16MB L2 cache anyone? 32MB? 64MB? Heck, why not just go for a GB? Who needs system memory anyway? -- that's a joke btw

Lastly, the group of people that need the best CPU on the market is rapidly shrinking. CPUs have gotten to the poitn currently where second and even third tier CPUs are good enough for most. Even the most punishing games today are GPU bound, not CPU bound. The only place that CPUs still matter is with businesses, and there power and scalability are more important. AMD still holds the crown there on both counts.

Re:AMD's new Power HOG

[Gr8Apes]

Pentium 4
Pentium D
Pentium M

You're right - crap is not good for consumers!

Re:AMD's new Power HOG

[rootmonkey]

Intel's new arch also has a true 128-bit FPU for SSE operations in 1 cycle as oppose to 2. Amd will have this with their quad core whenever that gets released.

Re:AMD's new Power HOG

[fitten]

Good points...

True SMP type multi-threading

What is "true SMP type multi-threading"? I've worked on parallel (multi-process and multi-threaded) code for almost 20 years now. There is no "one true" SMP processing type. The differentiators are data partitioning and how much serial vs. parallelism are inherent in your problem/algorithms. All are "true" in every sense of the word. Some are more efficient than others and some scale better... that's it.

As far as L2 cache sizes, they *are* going to grow with the 45nm shrink, it's already been announced... and that's fine by me as the 45nm shrink will just increase the frequency which means the gap between chip speed and memory speed will just go up. And yeah, for most things, I'd take a GB cache :) You'd still have to have main memory, even it were all cached, just because of the way that main memory and cache work differently (they are sort of the same but not - VM is sorta like a caching technique, but they operate on different levels so you have to have both still :p).

And yeah... the whole 'there'll only ever been a need for maybe seven computers' and '640k is enough for anybody' arguments have been around for a long time but have never held true. There'll always be a need for faster machines, but I do agree with you, a human can only send/receive email, surf the web, and type so fast and current CPUs are far beyond what it takes to do that. Still... as Valve has shown, build it and they will come... at least in the entertainment world (games and content), being able to have more CPU power is still good.

Re:AMD's new Power HOG

[pipatron]

3) Intel's architecture has already shown problems with the FSB limitations with Woodcrest in Apple benchmarks. True SMP type multi-threading causes degraded performance, especially when heavy large-scale memory access occurs that exceeds the cache capacity. I can't imagine this will get better @ 45nm, except that the cache might grow. 16MB L2 cache anyone? 32MB? 64MB? Heck, why not just go for a GB? Who needs system memory anyway? -- that's a joke btw

http://en.wikipedia.org/wiki/Cache_only_memory_arc hitecture

Re:AMD's new Power HOG

[Anonymous Coward]

Um, no. Core 2 Duo has more cache because it is based on 65nm process instead of 90. The transistors are smaller so they can fit more cache in the same space. And increasing the size of say L1 only barely affects the best-case latency because the cache is associative; you still only have say 2 or 4 or N entries in the cache that can hold a given address even after you double its size... you just have more space so can hold more lines in at once. Then you add in the savings of fewer misses and there's pretty much no reason other than cost to increase the cache size (ie how valuable the space is on the die).

Core 2 Duo is only winning right now because of the process, but Intel's trouble is that a processes relies on science and technology whereas a processor has to be designed. As time goes by it gets easier and easier for AMD to get to a certain nm, but it doesn't get any easier to design the CPU. And right now AMD's designs are much better than Core 2 Duo (they are in the same ballpark in performance/watt at .4x the density).

Re:AMD's new Power HOG

[fitten]

Core 2 Duo is only winning right now because of the process, but Intel's trouble is that a processes relies on science and technology whereas a processor has to be designed. As time goes by it gets easier and easier for AMD to get to a certain nm, but it doesn't get any easier to design the CPU. And right now AMD's designs are much better than Core 2 Duo (they are in the same ballpark in performance/watt at .4x the density).

How does this explain that when clock speed and L2 cache sizes are equal, Core2Duo outperforms Athlon X2 by a non-trivial amount? If it were "just process", then you could try to show where Core2Duo wins based on how much cache it has and the like, but that isn't what we see in the multitudes of benchmarks that have been run. A 65nm 2GHz part doesn't just magically perform calculations faster than the same part running at the same frequency at 90nm, for example.

Re:AMD's new Power HOG

[nbowman]

Err, no? Your article says: "The sample processors, known as the Penryn family within Intel, are being made at a factory in Oregon, and the company is on track to begin selling the chips in the second half of 2007, said Mark Bohr, director of process architecture and integration at Intel." and "AMD is expected to ship its first processors made using 65-nanometer technology by the end of 2006, and the company has said it wants to get 45-nanometer products on the market by the middle of 2008 as it tries to close the gap with Intel." looks to me like AMD will have shipped 65nm processors before Intel will be shipping 45nm according to the source you provided. even if AMD ships their first 65nm product in first half 07, that is still before Intel is slated to start selling their 45nm products.

Re:AMD's new Power HOG

[Svartalf]

They're not equal. They're not at all equal and won't be for a bit yet because of TDP, etc.

2Gb of L2 versus 1Gb of L2 per Core (1Gb twice over...) isn't the same thing. On non SMP benchmarks the 2Gb will win. Same goes with Four Cores and so forth. They don't do better on a "matched playing field" because it's not. Most of the 'benchmarks' don't really DO SMP operations- they're Uniproc applications, even the games... And on Uniproc, the DUO technically has double the cache because it's got 2Gb shared between the cores. Even with the dual core Athlon64's right at the moment, the biggest L2 they're fielding on the chips being compared is 1Gb per core. "Same memory" but effectively HALF the performance gain in those apps because they're pretty much running against one core and therefore have only 1Gb to work with, not mostly 2.

When you factor in true SMP operations, heavy computation, or the next generation of games even, the story becomes a little muddy.
Depending on what the mix of instructions are, you can end up with a stall on the L2 with the Intel part because the CPUs share that L2 and can be in different places at the same time- a cache spill's expensive on machine speed as you well know. The Intel part can conceivably be having a LOT of those under a real multiproc load. With that in mind, one CPU may have 1.5Gb of cache for it to work with reliably before a cache spill and 512Mb of cache on the other core; or even nastier combinations thereof. The Athlon architecture doesn't HAVE that concern- it's as if you've got two (or four) full Athlon64's working at the same time. With the Intel offerings, you're going to have swinging performance losses that you can't quite explain all the time. Sure, it's fast with what's now available, but it's not stable speed; bring in that second CPU into full play and things don't look as rosy.

Re:AMD's new Power HOG

[Superfarstucker]

I honestly think you pulled that straight out of your ass.

Re:AMD's new Power HOG

[fitten]

When you factor in true SMP operations,

Again... what is this mythical "true SMP operations" that people keep mentioning? Are you talking about MIMD code?

Depending on what the mix of instructions are, you can end up with a stall on the L2 with the Intel part because the CPUs share that L2 and can be in different places at the same time- a cache spill's expensive on machine speed as you well know.

I don't understand the "places" you mention. L2 cache has been multiported for a long time. Additionally, the cache subsystem should be able to handle simultaneous requests from both cores. There should be no stalling due to simultaneous cache accesses from both cores in a shared cache system. As far as cache spills, any situation that should cause spills in a shared cache should cause spills in non-shared (I'll mention this later). Basically, the shared 2M cache can mimic the degenerate case of two 1M caches exactly, but has the flexibility to also be the same as one core having a 512K cache and the other having a 1.5M cache, if working sets dictate, for example (I'll mention this later too).

The Intel part can conceivably be having a LOT of those under a real multiproc load. With that in mind, one CPU may have 1.5Gb of cache for it to work with reliably before a cache spill and 512Mb of cache on the other core; or even nastier combinations thereof. The Athlon architecture doesn't HAVE that concern- it's as if you've got two (or four) full Athlon64's working at the same time. With the Intel offerings, you're going to have swinging performance losses that you can't quite explain all the time. Sure, it's fast with what's now available, but it's not stable speed; bring in that second CPU into full play and things don't look as rosy.

I don't get your discussion... I'm just not following your verbage. I'm trying to understand it but can't get your metaphors or something.

Anyway I'll try to discuss what I think you are talking about. Shared L2 cache is considered the superior design compared to each core having unshared cache. There are numerous discussions on this around the 'net. However, I'll talk about several specific examples.

In a non-shared cache configuration with two cores on the same die running multithreaded code, you can easily get into situations where each thread wants access to the same piece of data for writing. When this happens (which is fairly common... mutex/semaphore/etc in fine-grained code are good examples of this), in a non-shared cache system, you can get a lot of MOESI traffic and passing around of that data between the two non-shared caches (takes inter-cache bandwidth to do that). However, in the shared cache system, that data is in the shared L2 cache exactly once and, furthermore, there is no passing it around... no MOESI traffic, no usage of any intercache bandwidth because no copy takes place. In such a situation (two threads competing for writes on the same data), the shared L2 cache can be very much faster than the non-shared L2 cache. In addition, the absence of the MOESI traffic is a lighter load on the MOESI subsystem, leaving it free to do other MOESI traffic and do other transfers. In some codes, MOESI traffic between non-shared cache and data copying between the unshared L2 caches can be almost pathological behaviour, leading to heavy slowdown as the two cores fight for access to the data. To summarize: Shared L2 = much lower MOESI traffic in a competing writes situation and little/no intercache bandwidth utilization because no copies between caches occurs. Non-shared L2 in such a situation is more MOESI traffic and intercache bandwidth utilized (and cores waiting for the data to transfer) to transfer the data back and forth. It's easy to write a simulation of this problem.

A second example is cache utilization. If you have two threads in a dual core system that are asymmetric in cache working set size, you can

Re:AMD's new Power HOG

[darkwhite]

A little less coherence and a little more Intel bashing and you'll sound like "Sharikou Ph.D"!

Re:AMD's new Power HOG

[Gr8Apes]

On the "true SMP type multi-threading", that would be multi-threading where the cores work on any thread, rather than assigning processes/threads to specific CPUs. This means that task1's work could occur in work units across any of the existing cores. This has a direct performance impact where task1 could run on core 1 with cache 1, then run on core 'x' with cache 'x', with the cache having most likely to be updated each time a task is run.

This is more related to how a task is run within the system than how you program your code.

I agree with your final point about faster machines eventually being needed. I see "agents" in the future that will parse your email etc for you, and act almost like a real assistant, perhaps eventually they will be a "real" assistant. But that's a while off, and for now, the CPUs are adequate. It'll be a year or two before quad cores will be "needed" for a game, not that they won't increase the experience, but they won't be needed. Much like the G-80 isn't needed to play Oblivion, but it increases the gameplay to the maximum possible.

required...

[jimstapleton]

I for one cant wait for our teraflop processing, cooking/water-heating AMInteD Core Optron Extreme Duo X2 overlords!

nforce chipset?

[Anonymous Coward]

it's strange that the new amd processor is only supported by an nvidia chipset, now that amd owns ati

Hotter?

[afidel]

The QX6700 has the same TDP(125-130W) per socket as the FX70-74 so I assume they run at about the same temperature on chip. Overall system temperature might be higher for the FX based quad core system since it uses twice as many sockets, but that's a matter of case design, if the case design can eliminate the heat from the heatsink effectively I would imagine both systems would run at the same temperature. This is of course ignoring the fact that AMD TDP is worst case and Intel's is average case.

Re:Hotter?

[traindirector]

I think the community of folks who follow processor awesomeness generally equate high TDP with system heat (and therefore ability to function as a space heater, cook eggs, burn the fingers of curious children, etc.) It's sort of sensationalist, in both senses of the term - "Man I bet you can feel the heat from that thing on the other side of the house!" Making fun of power draw just doesn't have as much good material.

Re:Hotter?

[Anonymous Coward]

The QX6700 uses only one socket; the FX-70 uses two dual-cores in two sockets, hence about double the power requirement.

Re:Hotter?

[Kjella]

The QX6700 has the same TDP(125-130W) per socket as the FX70-74 so I assume they run at about the same temperature on chip [so with enough cooling] both systems would run at the same temperature.

They might not run hotter (operating temp) but they produce twice the heat (power draw), so you're really pushing the semantics here. More heat = more cooling = more noise = higher power bill = lower battery life. These are what matters to end-users, I couldn't care less if it spends it all in one place or two.

Re:Hotter?

[jelle]

Nah, imho the socket itself is not the main reason for the difference. The qx6700 'processor' has two separate dies inside too, basically the same as the fx70 but then inside of the package. Sure, the traces are longer for the fx70 because it doesn't stay inside the package, but imho the main difference in power (besides the well-known difference in what the two manufacturers mean when they say 'tdp'), is the fact that the fx70 is built at 90nm and the qx6700 is built at 65nm.

Re:Hotter?

[MojoStan]

the FX-70 uses two dual-cores in two sockets

I think it's also worth noting that the QuadFX platform apparently doubles some parts of NVIDIA's power-hungry chipset (12 SATA ports??). Back when the single-CPU AM2 platform was launched, the NVIDIA chipset consumed a lot more power than the ATI chipset: somewhere between 20 watts and 40 watts.

Re:Hotter?

[MojoStan]

but that's a matter of case design, if the case design can eliminate the heat from the heatsink effectively I would imagine both systems would run at the same temperature

Which brings up another "con" for the QuadFX platform: so far, it's only available in the eATX (extended ATX) form factor. The motherboard is too big to fit in almost all popular gaming cases (which max out at "standard" ATX). In contrast, a Core 2 Quad can be used in standard ATX and even microATX SFF cases like the Falcon NW Fragbox.

I still don't understand the big need for the QuadFX platform when a dual-Opteron motherboard can be bought in the standard ATX form factor for less than $300 and Opteron 2000 series CPUs can be bought for as little as $208 a piece.

The race is on!

[Non-CleverNickName]

I wonder which will come first?

processors with 10 cores
or
razors with 10 blades

Re:The race is on!

[jdmicklos]

Razor blades?? Last I checked, razor blade numbers are not based on powers of 2...

Re:The race is on!

[Anonymous Coward]

And neither is the number 10

Re:The race is on!

[Anonymous Coward]

There are 10 kinds of people in the world, those that understand binary and those who don't.

Re:The race is on!

[c0nst]

there are 10 types of people in this world.. those who understand binary and those who don't ;)

Re:The race is on!

[dasunt]

I wonder which will come first?

processors with 10 cores
or
razors with 10 blades

Sparc is up to eight cores.

Razors have better hurry.

Re:The race is on!

[Gumby]

Well - the Cell wins!
8 SPU's plus a dual core PPC = 10 cores

Re:The race is on!

[leoc]

It's worse than that. Cell has 9 cores.

Re:The race is on!

[Anonymous Coward]

Gintelette. The best a geek can get.

Re:The race is on!

[Neoncow]

Psh! I know what category you're in. Did you realize that the article is about a 100 core processor?

Re:The race is on!

[Anonymous Coward]

Actually there are 2 kinds of people in the world, those who "don't" think that phrase is still funny (or 'l337') and those who "don't"...

Re:do they read what they write?

[Anonymous Coward]

How about reading the article after you write it?

In the process of writing the article, they read it as it is written in the process.

I for one...

[Abreu]

am still happy with my Duron 1300...

Re:I for one...

[Massacrifice]

You don't run Gentoo, do you?

Re:I for one...

[kalirion]

Well I'm satisfied with my P3-800, so there!

Re:I for one...

[CaptainNerdCave]

what do you do with all that processing power?

i'm barely maxxing out my intel 66mhz with the turbo activated.

Re:I for one...

[raynet]

And I am happy with my 1MHz Commodore 64...

Re:I for one...

[leoc]

I'm not very happy with my pile of rocks for counting, so I think I'm going to upgrade to an abacus once the price comes down just a LITTLE further.

Re:I for one...

[guaigean]

Then you probably don't do HPC. Yes, most desktop users are perfectly content with their clock speeds currently, and their single core systems (I am content with my home system), but in HPC, it's never enough. There are always bigger simulations, real time needs, etc. that will require more cores, more power, more flops.

Re:I for one...

[Sloppy]

What's really great, is that the longer you're happy with it, the awesomer your next computer will be, when you finally tire of this one. You won't be stuck with a mere 8-core 6 GHz box like all those other suckers who wasted their money on that obsolete crap back in 2009.

Re:I for one...

[Belial6]

I've been actually downgrading machines around the house. Epias with 1200MHz are just fine for 99% of the tasks, use less power and are completely silent. I'll keep my primary work machine as a power system, and I might have a medium level machine for my wife to game on, but that will usually be turned off.

Re:I for one...

[WuphonsReach]

You don't run Gentoo, do you?

The more up-to-date version would be:

You don't do virtualization, do you?

Start cramming multiple virtual servers onto a single box and all of a sudden dual-core solutions start to seem limiting. And you find yourself wondering just how much a 4-way quad-core machine would cost...

(That 4-CPU quad-core machine is still going to be cheaper then maintaining 4 separate quad-core servers.)

Corporate Stupidity

[eddy]

I think most here knew that this was always going to be a stupid vanity platform, almost as stupid as water-cooled memory modules. Now, the only thing more sad and stupid than a vanity platform, is one where the vanity isn't even there.

This should have ended as abandoned concept art in a drawer.

(PS. My current gaming rig is AMD X2-based, but if they don't have the performance/$ then they won't get in on the next upgrade)

Re:Corporate Stupidity

[RightSaidFred99]

Agreed. AMD should have butched up and just waited until K8L to take on Intel. This product is absolutely ridiculous, it's strictly for AMD fanbois with more money than sense. Let's see.. slower on almost every benchmark, very expensive motherboard and memory requirements, ridiculous power requirements, questionable future.

Yeah, it's a real winner.

Re:Corporate Stupidity

[dbIII]

I think most here knew that this was always going to be a stupid vanity platform

No - this will be the difference between having an 8 way system that takes up 5U in a rack and an 8 way system that takes up 1U since it only needs two sockets. You don't need a huge board that only goes into oversized cases - that said the huge board does have 16 memory slots.

Windows XP NUMA support

[Wonko the Sane]

The Extreme Tech benchmarks seemed to expose a lack of windows XP's ability to benefit from NUMA. I wonder what testing on a newer linux kernel with NUMA scheduler support would show.

Re:Windows XP NUMA support

[Mad Quacker]

This appears to be exactly the problem. The architecture/scalability is superior on paper, yet in benchmarks it's worse off than slower AMD processors. Take a look at the memory benchmarks. I bet XP is spending a lot of time moving threads between memory banks. All these tests are handicapped by inadequate OS support. Not great press for AMD, but I'm waiting to see what it will do on an OS that supports NUMA.

Re:Windows XP NUMA support

[Joe The Dragon]

TEST IT on VISTA 64

Re:Windows XP NUMA support

[kscguru]

Having done NUMA benchmarks ... on AMD chips, certain workloads take a latency hit from 60ns / memory access to 80ns / memory access. Bandwidth is halved. Net, it's a 5-10% slowdown across all workloads (5% if you try for average-case performance, 10% if you just hope for the best). Both sites point out that it is NUMA making the difference, yet both sites insisted on staying with Windows XP. A new motherboard like that, it's defaulted for a NUMA OS, so this is the 10% realm. As you point out, it would be extremely simple to run modern Linux (or Windows 2003, or Windows XP x64, or Vista) and see how well a NUMA scheduler works. (Note to Linux fans: Linux didn't have good NUMA scheduling either when Windows XP came out. A fair comparison would be against Linux 2.4.3 or so). This benchmark is fantastically stupid - it's the equivalent of running game benchmarks with a Voodoo3 graphics card to see CPU differences, determining that the graphics card is the bottleneck, then claiming one CPU is faster! Their benchmarks exposed a major slowdown in the memory system, easily corrected with an OS upgrade, and they refused to correct it.

In short, once you factor NUMA out of these benchmarks, the difference between AMD quad and Intel quad is approximately the same as the difference between AMD's K8 arch and Intel's Core arch for single cores. Umm... duh?

Re:Windows XP NUMA support

[Anonymous Coward]

Somebody mod this parent up. This one is very insightful!

HardOCP's take (link)

[Brian Stretch]

HardOCP QuadFX Review.

I'd go with the QuadFX platform just so I could swap in two quadcore AMD chips mid-2007, or one quadcore and one Torrenza platform coprocessor... if I had a few $thousand lying around and could make proper use of all that firepower. I suspect that quadcore + coprocessor combination is going to be really, really interesting within a year.

But can AMD gain on process?

[eddy]

Only problem then is that as it currently stands, Intel is, allegedly, ahead in the process game: "Intel to hit 65nm-45nm cross-over in 2008"

I like AMD for Hypertranspart, and Intel need to go there too sooner or later, but elegance isn't going to get AMD the win in the mass market; only performance can do that (and performance per dollar at that).

Sadly (for reasons of competition), I'm afraid that Intel may remain on top unless they run into problems with 45nm and AMD can sneak up on them by things going better than smooth.

As somewhat of an aside, maybe AMD could make some gains on features, so I still think they should implement a next generation PadLock (VIAs Crypto-Engine-On-CPU). I'd like for both AMD and Intel to go there together (compatible implementations), but one is better than none.

Once you've PadLock'ed, you can't go back.

AMD couldn't last forever on top

[arniebuteft]

But this will hopefully spur them on, to get back where they belong. I just hope they don't get desperate and start slashing prices - that may work in the short run, but long term, AMD will sink into "second tier" status, simply because they go cheap. AMD has surpassed Intel in quality for years, and has mostly shaken off the 'discount chip' label.

Nobody wants to be seen as the Dollar General of processors...Cyrix anyone?

Re:AMD couldn't last forever on top

[vision864]

Quality HAH!, compare the board situation between intel and amd chips -netburst crap included and seriously rethink that statement on the platform side amd has NEVER had near the kind of "quality"

Re:AMD couldn't last forever on top

[stunnet]

AMD is truly technology driven company. We can expect their product to get back to the top spot on a regular basis. AMD has the record for being the first 1 GHz processor, first 64 bit processor based on x86, first dual core processor... What intel is doing is following AMD and racing them using their muscle power. They have more fabs and more engineers to work for them if not better.

Check out AMD's wrongdoing!

[Anonymous Coward]

It's right here, folks: http://malfy.org/

AMD QuadFX

[Anonymous Coward]

Imagine a beowulf cluster of these!

Tom's Hardware...

[SirKron]

has another review that says reaffirms the same findings. Performance is not beating Intel yet and the AMD/ASUS solution is very expensive. I feel the only market here is those that cannot wait and have money to burn.

Quad cores - great for servers... but.

[NitroWolf]

Well, having lived with an AMD 64x2 for over a year now, I feel comfortable saying that a dual core proc is pretty useless to me. I've noticed absolutely no difference in my computing experience, either in the newest games or in day-to-day non-game activity. It's no different than my similarly clocked A64 with one core.

I'm sure quad cores are great for my servers, especially a couple of my mail servers that process a boatload of mail... but honestly, it's completely useless for the desktop. I would go so far as to say, at this time, and probably for the next couple years, dual core for the desktop/average user is completely useless. When and if games start using dual core, then it will be useful. Before that happens, only people doing heavy graphics work and other similar unusual (in terms of average users) computing needs will benefit.

All this BS about "Wow, I can burn a DVD and play a game at the same time" is just that... BS. Yeah, you might be able to do that, but who the hell cares? I've never had the urge to encode a DVD and play a game at the same time... why would anyone want to do this on a regular basis anyway?

Re:Quad cores - great for servers... but.

[geekoid]

I do.

I also like to play games and listen to mp3s. Game music has a tendency to get boring.
I also like to play games while downloading large files.

Many times I ahve been watching a DVD while burning a disk, or cimpiling and buring disks for distribution.

OTOH, I found that a SCSI 320 with 3 ms seek time takes care of this problem just fine.

Re:Quad cores - great for servers... but.

[N7DR]

Well, having lived with an AMD 64x2 for over a year now, I feel comfortable saying that a dual core proc is pretty useless to me. I've noticed absolutely no difference in my computing experience, either in the newest games or in day-to-day non-game activity. It's no different than my similarly clocked A64 with one core.

Stating the blindingly obvious: some people aren't going to notice much (if any) difference; others are going to see a huge difference. Parent falls into the former camp; I fall into the latter. I also have been using 62x2 for a year, and no way would I go back to single core. It would be worth having dual core if only for the fact that I can start a job and it will consume a core while all my interactive work runs on the second core, and hence I don't even notice that a huge job is running in the background. Everything else one gets with dual-core is an added bonus. I'm not totally certain that going to 4 cores on the desktop will be as useful, but I can believe that it might be, and will certainly be worth trying. For me, anyway (and I can't believe that I'm particularly untypical of slashdot users).

Given my experience, I'm even fairly convinced that the rest of my family (who are much more like ordinary users) would benefit from dual core too. Everything is simply so much more responsive.

Re:Quad cores - great for servers... but.

[Kjella]

Stating the blindingly obvious: some people aren't going to notice much (if any) difference; others are going to see a huge difference.

Given the number of PCs that were selling with faaaar too little memory arond when XP Home was released, I think most are oblivious. If they can't notice that their machine is swapping like crazy, what are they going to notice? Personally I think my biggest issue seems to be locking IO calls - the worst thing I can do to my Windows box is put in a semi-scratched CD/DVD. It'll spin up and down and finally read it while Windows locks up completely. You can tell the same is going on with heavy disk writes - everything else is sluggish like hell. I'm sure dual core helps - I just can't figure out why it needs to CPU lock like that in the first place.

Re:Quad cores - great for servers... but.

[Anonymous Coward]

"When and if games start using dual core, then it will be useful."

Huh? Most games of any real complexity these days will be multi-threaded, and thus will be using how every many cores it can. Think about it. Do you want rendering performance to degrade while something is being pulled off disk, or data moved through the network?

Game: "Ok, I'll be entering a warp portal in about 10 seconds, and will need to pull 50MB of textures from disk into memory. Let's fire up a thread to do that now, so the player won't see a long "loading..." screen."

Player's game runs on cpu0, while cpu1 pulls data off disks and does whatever in-memory formatting needs to be done.

Re:Quad cores - great for servers... but.

[reidconti]

Are you kidding me? I notice the performance difference all the time.

I spent 4 years on a 600mhz G3 iBook, and just upgraded to a Quad Mac Pro 2.66. I've also regularly used a P4 2.4 and a Celeron 1.7 laptop at work.

To start with, actually, I ran a 6-proc 200mhz PPro at an old job I had, and while the thing was never lightning fast, it ran like a train.. NOTHING slowed it down, ever.

I was continually surprised at how slow the P4 2.4 running XP felt, given my slow iBook at home... so part of it may be the OS.

And it's obvious to say that when I stepped up to the Mac Pro, it blew everything out of the water. But I can't believe you're saying multiple cores is only good for running multiple apps. That's simplistic and wrong. To begin with, we're ALL running multiple apps all the time -- MP3 player, web browser, email, and so on. Sure, none of these processes are taking a ton of CPU time, but the ability for the OS to assign them to different cores means your computing experience is much smoother and more consistant.

Hell, Acquisition X (P2P client) often has something like 50 java threads running, so even though I'm running "one program" its not like I'm not seeing a benefit from 4 cores.

Also, I am in the process of ripping my entire DVD collection to hard disk, and it's pretty sweet the way Handbrake can crunch a full DVD down to 1500kbps at full size/framerate in 40 minutes. I'm not using my machine as a server or as a professional, but that doesn't mean we don't all have needs for a high level of performance now and then.

The average person might not benefit a whole lot from more than 2 or 4 cores right now, but come on. The computing experience is SO much better, and it's not about running more than one intensive user application at once, it's about everything your OS runs, the multithreading of the application, and so on.

Now I just need FB-DIMMs to drop in price so I can get more RAM...

Re:Quad cores - great for servers... but.

[maraist]

I just can't figure out why it needs to CPU lock like that in the first place.

preface: I-am-not-a-windows-programmer...

My guess is that windows isn't actually locking, but instead that the file-system directory is locking... But since virtually any OS call is accessing some kind of file object, those OS calls will likely lock too.

Same is true in UNIX, except UNIX isn't retarded enough to have unrelated virtual paths block one another. Gnome, on the other hand... Well it does still dream about becoming windows/Mac one day..

The key problem as I see it with windows is the super-directory. Most directories have 2 or 3 levels of indirections to them that absolutely necessitate being in sync with one another.. 'My documents', 'Temporary [internet] files', 'Shared folders', hell, even the drive letters. I assume that the definition of a path therefore requires a central lock. And something like the addition / removal of a device requires the locking of this directory.

If you had an OS file number which directly mapped to an inode, which directly mapped to an immuteable partition somewhere, then I couldn't imagine there being a need to lock along with the central directory.. But if your file accesses are always of the form "J:\My Folder\blalba" with two levels of virtual/dynamic directories, then good luck.

Note, I'm hoping that subsequent accesses to an already opened file don't block, but who knows.

Re:Quad cores - great for servers... but.

[NitroWolf]

And it's obvious to say that when I stepped up to the Mac Pro, it blew everything out of the water. But I can't believe you're saying multiple cores is only good for running multiple apps. That's simplistic and wrong. To begin with, we're ALL running multiple apps all the time -- MP3 player, web browser, email, and so on. Sure, none of these processes are taking a ton of CPU time, but the ability for the OS to assign them to different cores means your computing experience is much smoother and more consistant.

Every application you mentioned doesn't consume enough CPU time to make a whit of difference to the average user in terms of single or multiple cores. It's laughable that you say "It means your computing experience is much smoother and more consistent," you are just repeating marketing bullshit. You've heard it so often that you believe it. I truly believe that YOU believe it's the case, but the fact is... it's not. If you have a fast single core CPU and you are doing the above applications like an average user, you aren't going to have a "smoother" experience. Your experience will be identical. Only pedantic nit-picking will reveal a difference. In a day-to-day basis, on average, there will be no difference.

Hell, Acquisition X (P2P client) often has something like 50 java threads running, so even though I'm running "one program" its not like I'm not seeing a benefit from 4 cores.

Tell me how you are seeing a benefit? I'm speaking to Windows and multi-core procs. On a Windows box, something with 50 threads is going to behave just the same to the user if they are on single or dual cores. I've had no experience with Macs since 1995, so I can't speak to them, and frankly I don't care. Macs are useless to me - they have and do nothing I want, so I have no interest in them. What your experience on a Mac with single or Dual Cores is, is totally irrelevant to the majority of computer users. You are a minority, live with it.

Also, I am in the process of ripping my entire DVD collection to hard disk, and it's pretty sweet the way Handbrake can crunch a full DVD down to 1500kbps at full size/framerate in 40 minutes. I'm not using my machine as a server or as a professional, but that doesn't mean we don't all have needs for a high level of performance now and then.

So you do this on a daily basis? Once you're done ripping your entire DVD collection to hard disc, you erase it all and start over? Oh... you meant it's a one time thing, didn't you? You are going to encode them once and then you're done. Well, golly... it's certainly worth plunking out $1000 for a quad core to rip your entire DVD collection, then after you're done 3 of those cores can go to waste! I mean... really. If that's what you want to do, why not just buy a cheap ass Dell, put it in the corner and let it rip your entire collection. Then when you're done, at least you can sell it on eBay or something.

Go read the original post before replying with nonsense.

Re:Quad cores - great for servers... but.

[NitroWolf]

Given my experience, I'm even fairly convinced that the rest of my family (who are much more like ordinary users) would benefit from dual core too. Everything is simply so much more responsive.

What? What is more responsive? Someone show me some tangible proof that day-to-day activities are "more responsive" or "smoother." I have yet to see it. Maybe I just set my shit up properly from the get-go, so my computing experience is already hyper responsive, so I don't notice the difference. Maybe dual and quad cores are good for the masses that have 500 task bar thingies running, a bunch of spyware and other cruft clogging up the system. Hell, if that's the case, then I take it back... dual and quad cores will be a boon to the average user... but totally useless to the competent user who doesn't need to compile shit in the background or rip DVDs every night while playing Oblivion.

Maybe YOU have a bunch of crap running that you don't need to have running, making your system sluggish and you see a benefit from multiple cores. Hell, I invite anyone to work on my system(s) for a day... I won't tell you which one is single core and which one is dual core... see if you can guess. They are both equally responsive.

Stupid

[sigmapsicharlie]

Man I love AMD and I wish they would make better decisions, when it comes to their chip design. They can not expect to compete with Intel when they use an outdated process technology of 90 nanometers, when Intel is using a smaller and more efficient 65 nanometers. I remember when AMD would run all over Intel's numbers, now AMD is losing the edge that gave them that big contract with Dell. If they would have actually tried a little harder and made a smaller faster chip they may have had a chance of knocking Intel out of the "throne", but now with these new processors it seems that is not going to happen. Best of luck to ya AMD, I hope you get back in the game eventually.

These are still 90nm chips...

[dlenmn]

So it's more of the same, no wonder it's not so impressive. Once they get 65nm stuff out, we may see real improvements (not only speed, but power consumption too).

Re:These are still 90nm chips...

[gyranthir]

http://www.tomshardware.com/2006/11/30/brute_force _quad_cores/ Read the Tom's article they explain the 90nm process and why these are more advanced than the intel chips. Gotta remember this is revision 1 of this system, we have yet to see how far this can be taken.

Re:These are still 90nm chips...

[dlenmn]

The article you linked to says, "AMD's current 90 nm silicon-on-insulator process is more advanced than the Intel equivalent". So AMD's 90nm process is more advanced that Intel's 90nm process (what else could they mean by "equivalent"?), but Intel is using a 65nm process and is gearing up for a 45nm process. What the article did NOT say (because it's not true) is that AMD's 90nm process is better than Intel's smaller processes, which is what I think that you're claiming.

I still think that AMD's move to 65nm should be more impressive.

Re:These are still 90nm chips...

[stevel]

The article does NOT "explain the 90nm process and why these are more advanced than the Intel chips". Rather, it just makes this bald assertion and moves on. Intel, I'm sure, would disagree. AMD and Intel use somewhat different silicon technology, but is one "more advanced" than the other? Depends on whom you ask. What AMD solves with SOI Intel solves a different way. There is, of course, the fact that Intel has been cranking out very successful 65nm chips for about a year now and AMD has not (yet) shipped any.

I don't think anyone can reasonably argue that Intel's process technology is behind AMD's. Consider "overclockability" - Intel's 90nm chips have been proven to have lots of headroom (think Pentium D 805) and the AMD chips fall over with even slight overclocks. This says to me that Intel's silicon design is inherently better. Whatever Intel's past failings have been, and there have been many, silicon process technology is not among them.

NUMA NUMA support

[Anonymous Coward]

http://www.youtube.com/watch?v=60og9gwKh1o

If it don't fit...

[Anonymous Coward]

If it doesn't fit in socket 939, we don't want it...

Tired of all of them unable to standardize on one processor socket. How many gd different sockets are there out in the field now... I lost count.

Check out the Tom's Article

[gyranthir]

Before you pass judgment on this. This is Rev 1, let's see how far they can take this. http://www.tomshardware.com/2006/11/30/brute_force _quad_cores/

Depends on what you call a "power user"

[Moraelin]

... if by "power user" you mean "someone who uses lots of power," then yes.

Here it comes

[geekoid]

the "but but but it's only ...." Fanboy responses.

When AMD comeones out with something thats better, then compare to whats out. not planning to be out, or njot quite ou, but actually available to the average consumer.

print url

[porsche922]

http://www.hothardware.com/printarticle.aspx?artic leid=911

Cores versus speed increase

[Anonymous Coward]

they are adding more and more cores, but the CPU speed is still what it was 3 or 4 years ago.
I'm supposed to be at 10 GHz now.

If I can't have my flying car, at least give me my 10 GHz CPU.

Misnomer?

[Anonymous Coward]

Two sockets! Four cores!

AMD's long rumored Dual-DualFX! Code named 2x4!

For some reason this particular stopgap in the multicore battle just seems really, really childish.

Same Price for Intel and AMD?

[fljmayer]

I don't understand why they say the prices for Intel and AMD quad core system are the same when the Intel QX6700 seems to go for $1500 on newegg, while a pair of AMD CPUs seems to range from $600 to $1000 (couldn't find real prices yet).

Re:do they read what they write?

[Anonymous Coward]

>>How about reading the article after you write it?

>They're in the process of reading it now.

Nah, we are not. We prefer to read slashdot instead. And here nobody RTFA - exept for this werido grandpa. He is probably new here and was rightfuly modded down.

Editor.

Re:L1 vs L2 vs RAM?

[WuphonsReach]

So... L2 cache speed. When I look at Memtest86+ numbers, I see:

~19700 MB/s for L1
~4700 MB/s for L2
~3000 MB/s for main memory

This is on a Athlon64 X2 4600+ w/ low-speed DDR2 RAM (4 sticks of 1GB).

I'm guessing that L2 gains are because it can respond to a memory request faster (fewer clock cycles) then because of the bandwidth? Because the L2 bandwidth of 4.7GB/s doesn't seem to be that exciting anymore once main RAM can feed the CPU at 3GB/s.

What is the reason for this much power?

[Sneakernets]

Unless you're running Dreamworks studios in your basement or running a simulation for MIT, I don't see the usefulness of this.

Show me the SPEC numbers!

[Phatmanotoo]

ExtremeTech has a plethora of application and synthetic benchmarks on QuadFX, including gaming and media-encoding tests."

Bleh. The benchmarks I'd really like to see are the ones geared towards scientific computing, like STREAMS and SPEC . Nowadays the Intel chips seem to score better on those "gaming" and "media-encoding" benchmarks, but that doesn't neccessarily mean that Intel FPUs are better for general scientic computing. (in fact, experience so far shows the opposite)

Re:Show me the SPEC numbers!

[AnInkle]

Check out The Tech Report's review, if it's scientific real-world benchmarking you want.

I'm wrong above

[dbIII]

OK - I'm wrong after reading the fine article - this is a 4 way dual CPU system and not the 4 core chip I thought it was from the name. The 5U system descibed above is 4 dual core AMD Opterons set up to be 8 way on a ridiculously big board that I hope to have very soon. What would be very nice is a two socket system with 4 cores per socket that can fit into 1U in a rack.

Any games out there on linux that would really show off an 8 way system to "test" it over the weekend? Whatever happened to the cluster quake code - did it get released?

AMD's FPU

[DrYak]

It's interesting that you mention the FPU, because I'm still thinking that there are some nice thing that could come out of AMD's Fusion project.

Fusion may only be a CPU and a graphic card in the same package, targeted at the entry market, as some have speculated.
But it can also turn out to be something more similar to the Cell processor a CPU with several general purpose Vector units that could be used for higher level computation (Physics, Geometry, or even scientific calculation @Home) while leaving the grunt work to the GFX cards (polygons and video decompression)

If AMD manage to bring out such a chip in the near future they'll be able, once again to surprise the market and attract interest in some fields (hard-core players, scientists, ...) even if they don't have intel's latest shrinking technology.

(Just like they managed to capture interest back with their 64bits extensions)

Re:AMD's FPU

[fitten]

Maybe... Microsoft's DirectX10 has an API in it for offloading vector type work to 'something else' in the system. The interesting thing about it is that it will be a standard API, meaning that hardware can be built to take advantage of it while drivers can also be written to either do it in emulation or by actually handing it off to the specialized hardware. This would help AMD out a lot as far as that kind of hardware goes... without some standard APIs, it would likely end up in a mess, IMO.

Personally, I'm not that excited about that kind of technology just yet because it is still fairly immature as far as PCs go. The logistics are rough all the way around. First hardware can likely be surpassed quickly with newer/better coprocessors but you have to a) replace the entire CPU, b) leave that coprocessor there but disable it or something in preference to an add-in card that is better (just like embedded graphics today) which means you have basically a dead 'core' on your CPU, and c) AMD will be stuck with a ton of dead stock as soon as they upgrade that coprocessor and AMD already has problems with keeping channels fed. c) will probably mean that advancement of Fusion will be slow. If AMD can release a Fusion vector part that is 2X as fast or has a better API (new version) or something, but the 'main core' is the same (say it's a 3.4GHz Athlon64 core) because it hasn't advanced as fast, the instant the new part comes out, no one will want the old one and it will just sit there unless it has huge discount. IMO, this will make AMD not want to advance the Fusion device at a faster pace than the x86-64 core that's also on the die because it would be very costly.

Re:Windows XP NUMA support (can blow me)

[billhuey]

Anybody running a 2.4.2 version of the Linux kernel should be shot. Nobody runs 2.4.2 these days and anybody suggesting that is far out of touch with what Linux is doing. Compare it against 2.6.19 with all of the NUMA options turned on (CPU local memory allocators, RCUed algorithms) and you'll see an expected an expected trumping of XP for kernel load hands down because of all of the MP work on it over the 4 years.

Why wait?

[Anonymous Coward]

I'd go with the QuadFX platform just so I could swap in two quadcore AMD chips mid-2007, or one quadcore and one Torrenza platform coprocessor... if I had a few $thousand lying around and could make proper use of all that firepower. I suspect that quadcore + coprocessor combination is going to be really, really interesting within a year.

Dual-socket Xeon platforms, which can accept Clovertown, are available today.
By next week, you could have 8 cores in your box!

It's WinXP stupid!

[Anonymous Coward]

Almost ALL the reviews have missed one crucial thing: WinXP sucks at NUMA and AMD uses NUMA (vs Intels SMP) memory architecture. When this is taken into consideration (i.e., VISTA), the results tip in AMD's favor. Try reading Charlie Demerjian's article in the INQ: http://theinquirer.net/default.aspx?article=36087

---John

Re:L1 vs L2 vs RAM?

[darkwhite]

Yes, L2's main benefit is latency, it should be almost 2 orders of magnitude faster