Details of Intel 45nm Processors Leaked

DCC writes "TechARP has gotten some juicy news from Intel. This time, it's the top secret details of the Intel 45nm desktop processors, both Yorkfield and Wolfdale with benchmarks and pricing included! 'As promised earlier, Intel will launch their 45 nm processors by the end of this year. In fact, we have been told that the launch date had already been set at November 11, 2007, so mark your calendars. [...] Code-named Yorkfield XE, the Intel Core 2 Extreme QX9650 will be a quad-core processor built from two 45 nm Wolfdale processor dies. It will displace the Core 2 Extreme QX6850 (Kentsfield) processor as the top desktop processor model until Q3, 2008'"

Comments

Not all that new

[CajunArson]

Anandtech had a preview of Wolfdale including benchmarks back in August (here). The ironic thing is that with the limited availability of the K10 and its late arrival at most review sites, I've seen about as much real benchmarking of the unreleased Intel parts as I have of the supposedly widely-released AMD parts.

Re:Not all that new

[mdm-adph]

I wasn't even aware the K10 was out, thus far I've seen nothing on it on any of the sites I normally look at. Very strange.
I know -- it's been a weird release, to say the least. I haven't really heard very much about them at all, and for a chip this neat that's kind of surprising.

You can buy them now on Newegg here -- they were up a few weeks ago for about $800, but then they were taken down, and now they're back up. Who knows, eh?

Time for a new naming convention?

[Psychor]

Maybe it's time to come up with processor names that actually mean something again instead of confusing and usually meaningless numbers? This is especially true for AMD, whose numbers seem to be based around the clock speed an equivalent Intel chip might have run at many years ago when they invented the convention, but Intel's new "random model numbers" naming doesn't seem much better.

Of course old style Megahertz numbering doesn't make much sense these days either, with the proliferation of multi-core processors. I think it would be nice if the chip makers could agree on some kind of general performance benchmark number that could be used in names to make processors more easily comparable. Even some kind of very basic number relating to cores/speed like the 4x2200 for a 4 core, 2.2Ghz chip would be better than the current mess in my opinion though.

Re:Time for a new naming convention?

[EvanED]

Even some kind of very basic number relating to cores/speed like the 4x2200 for a 4 core, 2.2Ghz chip

Okay, now how do you mark different versions of that? Ones with different sized caches? Different FSB speeds?

I'm not claiming that the Intel numbers make all that much sense, but they still manage to convey a fair bit of information. Higher "hundreds" digits are faster clocks. (The Q6600 and E6600 both have the same clock speed.) Numbers with the same leading digits, e.g. the E6700 vs. the E6750, are different revisions. (The E6700 has a 1066 MHz FSB, and the E6750 1333 MHz.) The E prefix says that it is dual core; quad cores have Q. If the thousands digit is 2, then things shift around a bit, but that information alone tells you that you're working with one of the budget chips. (Slower clock, no VT, smaller cache.)

Re:Time for a new naming convention?

[qortra]

some kind of very basic number relating to cores/speed like the 4x2200 for a 4 core, 2.2Ghz chip
Of course, that would insufficient; You would need some other indicator to mention that fact that it is, say, a Wolfdale instead of a Conroe (Wolfdale's being, say, 10% faster). Also imagine that another axis has to be considered; power-efficient, or non-power-efficient. That would make your model name even more complicated: "Conroe4x2200PE". That's quite a mouthful. This is only an example to indicate that specification-based model numbers have a tendency to get prohibitavely complex. The spec-based model number becomes more and more complicated as each axis of of variability is added. So, most companies settle for more arbitrary model numbers (NVidia, ATI, BMW, Lexus, etc).

Consider BMW; for many years, their model numbers were usually of the form *model**displacement**feature* (an incomplete schema would be [3,5,7][15,20,25,30,35,40,50][i,ci]). However, at some point, they realized that there might be several different models with the same engine size, and rather than making the feature indicator more complex, they made displacement inaccurate (sometimes pro-rating it to indicated a higher performance version of an identical displacement engine).

All that to say, I don't think that having arbitrary model number is such a bad state of affairs; just have a publically accessible lookup table that relates a single model number to all the relevant specs. Saying "NVidia 7600GT" is a heck of a lot easier than saying "NVidia 256-128-22.4-6.7-700" (for memory, memory interface, memory bandwidth, fill rate, and vertices-per-second).

Re:Time for a new naming convention?

[TheThiefMaster]

For the AMD Athlon 64 X2 processors, the number does actually mean something.

For a 1MB cache (per core) cpu, it's exactly 2x the clock speed in megahertz. The X2 4000+ is 2000MHz. This continues every 200MHz all the way up to the top cpu, the 6400+ (3200MHz, 1MB cache).

For a 512kB cache (per core) cpu, it's 200 lower than that. The X2 3800+ is 2000MHz as well, but 512kB cache. This continues every 100MHz all the way up the line to the 5400+ (2800MHz, 512kB cache).

For a 256kB cache (per core) cpu, it's 200 lower again. The X2 3600+ is ALSO 2000MHz, but has 256kB cache. There is only one 256kB cache X2 cpu. There is also a X2 3600+ that is 1900MHz and 512kB cache, which still fits the pattern.

The single core Athlon 64s seem to have a similar numbering scheme, but with more factors affecting it, including hypertransport speed (800MHz/1000MHz), and socket (754/939). Some of the cpus were numbered slightly differently, but this is 99% accurate:
The base is a 512kB cache socket 754 hypertransport 800MHz 2000MHz cpu, which is rated at 3000.
Socket 754 cpus were rated 200 higher for every 200MHz higher cpu speed.
1MB cache versions were mostly 200 higher (one was 300), and 256kB cache versions were 100 lower.

Socket 939 cpus were rated 200 higher than socket 754. (Due to the support for dual-channel ddr, they were better).
1000MHz HT cpus were rated an additional 100 higher for every 200MHz higher cpu speed than the base (2000MHz). The cpus that were 200MHz slower than the base didn't get an additional 100 points deducted though.
Again, 1MB cache versions were 200 higher.

This doesn't cover the 1500+, which was only used in a HP Blade PC.

The AM2 cpus were mostly the same as the 1000MHz HT S939s, except for the 4000+, which was a 2600MHz/512kB cache instead of 2400MHz/1MB, and the details above would have scored it at 4100+.

As you can see, the numbers are mostly arbitrary, and mostly derived from the features of the cpus instead of a comparison against intel.

Re:Time for a new naming convention?

[dextromulous]

I think it would be nice if the chip makers could agree on some kind of general performance benchmark number that could be used in names to make processors more easily comparable.

You mean like BogoMIPS? That reminds me of my favorite lines from the Linux 2.4 source code: arch/i386/kernel/smpboot.c
/*
* Allow the user to impress friends.
*/
After which is the calculation for bogomips.

Still FSB and dual dual-core

[Joe The Dragon]

The true AMD quad-cores may blow intel away the desktop ones will use faster and lower lag desktop ram then the sever ones that are out now.

And The amd 4x4 system with 2 amd quad cores with desktop ram will be alot better then intel Skulltrail with FB-DIMMS and poor chipset io Full sever chipset + 2 nvidia chipset linked by a pci-e x16 bus 1.1 from the intel chipset to the nvidia chip and HT from nvidia to the other nvidia chipset with 2 x16 pci-e 1.1 sli slots. Amd system will cost less with cheaper ram and
a less costly MB.

The amd system will likey have the choice of a nvidia based system with 2 Full sli x16 slots pci-e 2.0 slots + other pci-e 2.0 slots with HT links form the cpus to the nvidia or a
ATI one with
        * Codenamed RD790
        * Dual or single AMD CPU configuration
        * Supports socket AM2+ and socket F CPU
        * Allowing maximum of four physical PCI-E x16 slots at x8 lanes bandwidth or 2 PCI-E x16 slots at maximum bandwidth (16x-16x or 8x-8x-8x-8x CrossFire)
        * Discrete PCI-E x4 slot
        * Providing a total of 52 PCI-E lanes [4], 41 lanes in Northbridge
        * Two to four cards CrossFire, with reported 2.6 times of performance than single card
        * Support of HyperTransport 3.0
        * Support for HTX slots
        * Support of PCI-E 2.0
        * Supports Dual Gigabit Ethernet, and teaming option
        * Discrete chipset cache memory of at least 16 KB to reduce the latencies and increase the bandwidth
        * Reference board codenamed "Wahoo" for dual-processor (Quad FX) reference design board with three physical PCI-E x16 slots, and "HammerHead" for single socket reference design board with four physical PCI-E x16 slots, also notable was the reference boards includes two ATA ports and only four SATA 3.0 Gbit/s ports (as being paired with SB600 southbridge), but the final product with SB700 southbridge (see below) should support up to six.
        * Northbridge runs at 3 W when idle, and maximum 10 W under load

http://en.wikipedia.org/wiki/AMD_700_chipset_series

Re:Still FSB and dual dual-core

[CajunArson]

Every single review I saw of the 4x4 had it losing to Intel quad cores using the "crippled" FSB. Hypertransport is great for 4 socket+ systems which is why Intel is going to a point-to-point interconnect next year. However, on smaller systems like desktops and up to 2 socket servers, Hypertransport's benefits are much less clear. For example, when Anandtech did the initial K10 benchmarks it turns out that it took the K10 about 76ns to transfer data between any 2 cores on the chip using its layer 3 shared cache (which is faster than the Hypertransport used in the 4x4).
    However the more interesting number was that it took Intel's FSB 77ns to transfer data between the dual-dies, and if the data were only going between cores on the same die that time was only 26ns. So the upshot was, that the worst case scenario for Intel's data latency was less than 2%, while the better case scenario (which is not too hard to achieve) gave Intel a 50% reduction in data latency. If you want to talk about 4 socket+ systems then Hypertransport is a winner, but on a desktop I wouldn't obsess over it too much.

Re:Still FSB and dual dual-core

[RightSaidFred99]

Skulltrail and 4x4 are for bragging rights only - pretty much nobody will buy either except a few uber-rich people who don't care about money. That said - the numbers will belie your impression. Skulltrail will score almost identically to 4x4 because it's all about SLI/Crossfire. In terms of raw computing power, Skulltrail will be superior to Barcelona simply because Barcelona, which is a good core, will only ship at 2.0GHz or _maybe_ 2.5GHz this year.

This really isn't a good time to be an AMD fanboy, I'm afraid - not like a few years ago when their products were better in pretty much every way than Intel's.

Re:Still FSB and dual dual-core

[TeknoHog]

I don't get this alleged superiority of "true" quad-cores vs. dual dual. I used to think that two discrete CPUs should outperform a dual-core, because the latter must share the external communication channels, other things being equal. Now I'm not so sure of it, because of things like shared cache and hypertransport, which may improve performance in certain situations, but it's not obvious. The difference between a dual dual and a quad is much more subtle. Would you prefer a "true" octal-core to your dual-quad contraption?

I thought it was the internal chip specs leaked..

[Marcion]

... but it turns out to be some pricing details.

Nothing to see here, move right along.

Was This An Accident?

[Nom du Keyboard]

Was this an accident, or FUD to put the brakes on AMD sales prior to the official release of the Intel processors? This way Intel get the news out without collecting the grief that comes from pre-announcing their next moves themselves.

here we go again

[Connie_Lingus]

Jimbob's Corollary to Moore's Law...

Every 18 months I will become ever more numbed by the announcement of denser and denser chips.

Re:here we go again

[644bd346996]

Not likely. Intel is currently developing their 32nm technology, and IBM has tested 29.9nm lithography. That's only around 600 times the Bohr radius (radius of a hydrogen atom). Within the next 10 years or so, we will have reached the fundamental limits on the size of a silicon transistor, and once those chips are brought to market, that's it. If Moore's law continues at all, it will be applied to something like quantum computers, not semiconductors.

Of course, there are many parts of a CPU that traditionally don't scale as well as the basic transistor, so with continued work, we can probably keep shrinking CPUs. But we'll be doing it in small increments with increasing marginal costs, not by the factors of 2 we've been seeing for the past 20 years.

Mod Parent Informative (At least)

[mpapet]

This kind of summary is a pet peeve of mine. "Top Secret Whatever is Leaked!" like this is advertising disguised as news.

Given the end-of-year release of the product, it's in sales, marketing and mass production hands now so there's nothing secret about it.

As a general rule, if something is "leaked" 3 months out, then it's advertising disguised as news because the product is ready for market, sales reps are out placing & promoting the product.

Parent is right on.

Leopard?

[Anonymous Coward]

I wonder if the processors will be announced in the form of spanking new Mac Pro towers to coincide with the release of Apple's Leopard? It's the kind of big glitzy media event that Apple love and Intel would love to be included in.

Flawed Analysis

[paulsnx2]

In the article, the author scaled the performance based on the clock speed each time a comparison was made between chips with different clock speeds. This was mostly done in favor of the new Intel chips.

The problem is right there in the Author's analysis. For example:

"If you extrapolate the data, then the Yorkfield processor is really about 12-21% faster than the Kentsfield at the same clock speed. This is almost entirely due to the 50% larger cache in the Yorkfield processor. The very large 81% boost in DivX 6.6.1 is again mostly due to SSE4-optimized code in DivX."

But But But!!!! Changing the clock speed doesn't make the cache any bigger! You can't then assume a linear relationship between performance and clockspeed if the difference is primarily how long you are going to have to wait to fill the cache!

The article isn't too flawed. They give actual results. But do yourself a favor as you read the aritcle and completely discount any "extrapolation" done by the author to get "really" numbers. When comparing processors, the "really" numbers are always the hard cold facts, not the "I wish" numbers generated by speculating what would happen if you changed the processors in some way.

Re:Yep

[Chirs]

I suspect it's just the principle of the thing. Intel is calling it a quad-core, but us techno-types know that it isn't. AMD's really is quad-core.

There is a perception that AMD's solution is more elegent and regardless of benchmarks is somehow "better". Intel's design is a "throw cache at the problem" sort of solution--but it works for most normal usage.

I suspect that many people would like to see what Intel could do if they got off their seats and really did something original...like if they can do this good with half-assed lashups, how good could their cpus be if they actually did some novel *design*!?!

Better value than a moderatly OCed Q6600 SLACR?

[funkdancer]

Just wondering if anyone here thinks the new CPUs will deliver better value than the Q6600 SLACR, which only costs around ~A$350 (US$315) and will easily reach 3ghz with virtually nill effort. Put a little bit of work in it and it will reach 3.4ghz or even 3.6ghz with just a little bit more, on a good air cooler

Whilst the running costs would be lower due to the lower energy usage, I'm just wondering if any of the new CPUs will come anywhere close to the absolutely fantastic performance/value that is currently represented by the SLACR.

I'm looking to buy a new CPU & motherboard for my Zalman HD160XT HTPC case in the next month or so. I already have a Q6600@3GHZ in my self built desktop (based on Asus Blitz Formula in an Antec Nine Hundred w/2GB of RAM) and it is supreme in desktop usage with lots of apps running in Vista, just totally outclassing the Core2Duo 2.67ghz WinXP desktop (IBM IntelliStation M Pro 9229/also 2GB of RAM) which I have at work.

Re:Yep

[therufus]

LOL, mod parent up funny!!!

Wait... you mean that wasn't sarcasm?

Look, to start with I'm an AMD fanboy (I guess), mainly due to the fact they call a spade a spade and don't lie to their (marketing zombie) customers about what their chips actually are. Intel are a marketing company first, a CPU manufacturer second. If you want to believe that the Q-series CPU's from Intel are actually quad core, you can take your ill-informed self to your nearest retailer and buy your double-duel-core CPU with your hard earned cash. Thats what your master... I mean... Intel want you to do.

But 4 cores on the one chip, a quad core does not make!

As long as data has to jump on the bus and take a trip down pipeline lane in order to see the neighbour next door, it's NOT A QUAD CORE! If you were to classify it, it's a 4-core CPU. With AMD's design, all 4 cores have equal access to data and can split and share properly.

There are even some applications out there that run SLOWER on an Intel 'quad core' CPU due to the constant bus hopping going on.

Why can't AMD 'keep up' as you say? Well, it's simple really. Because Intel market the crap out of an average product, they sell more units based on lying and deceiving customers. Not to mention the kickbacks and underhand tactics with your tier one OEM's (Dell, HP, etc). Sell more CPU's and you'll get more money. AMD have taken the noble road. They're suffering for it now, but they'll be on top again, just as they were with the first 1000MHz CPU (beat Intel to that cherry) and the first 64-bit desktop CPU (oops, Intel lucked out again). Come to think of it, the Athlon X2 was out before the Pentium D if I recall correctly (that seems to be the trifecta).

(sarcasm)Oh yea, AMD sure can't think things up before Intel.(/sarcasm)