Original Link: http://www.anandtech.com/show/1711



Anand and I had a discussion about a year ago that went something along the lines of "2005 is going to be sooooo boring for Intel and AMD". But then a week later, AMD and Intel both started announcing dual core stuff internally and 2005 has been one of the better processor slugfests that we have seen yet. After passing up the last few roadmaps due to various tradeshows and NDAs, we decided to plunge into the most recent Intel roadmaps and really take an in-depth look at what is coming up for the next year or so.

We have talked extensively about Yonah - the dual core successor to Dothan - without much attention to its clock speed and other features. The latest roadmap has provided quite a bit of new information about several of the cores. We'll take a more in-depth look at Yonah, Presler and Cedar Mill in this update.

Desktop

Over the last month we were bombarded by dual core announcements and releases from AMD and Intel. Not only because dual core is an interesting and new concept to desktop computing, but also because after all of the hype, AMD and Intel were both able to deliver fairly good products (both AMD [RTPE: AMD Opteron Italy] and Intel [RTPE: "Pentium D"] are shipping already). Inevitably, there have been some pretty major changes to Intel's product naming, for starters:
  • Pentium 4 Extreme Edition processors are now simply named Pentium Extreme Edition. Although these processors are nothing more than best of breed server and desktop cores, they will now have their own product category at least. (This started with the most recent dual core Pentium XE.)
  • Pentium 4 now only refers to the single core Prescott and upcoming 65nm Cedar Mill cores. Pentium D will refer to Smithfield and 65nm Presler.
  • All 5x1 processors are now 64-bit enabled.
  • All 6x2 processors are now VT enabled.
  • All 6x3 processors are 65nm Cedar Mill (which has VT enabled).
Let's take a look at the upcoming dual core roadmap.

Intel Dual Core Performance Desktop Lineup LGA775

Processor

Speed

L2 Cache

FSB

Launch

Pentium D 950

3.40GHz

2x2MB

800MHz

Q1'06

Pentium D 940

3.20GHz

2x2MB

800MHz

Q1'06

Pentium D 930

3.00GHz

2x2MB

800MHz

Q1'06

Pentium D 920

2.80GHz

2x2MB

800MHz

Q1'06

Pentium D 840

3.20GHz

2x1MB

800MHz

Now

Pentium D 830

3.00GHz

2x1MB

800MHz

Now

Pentium D 820

2.80GHz

2x1MB

800MHz

Now


Aside from the additional cache, also note that the new 9xx series processors are the same 65nm Presler cores mentioned earlier. Presler will in fact be nothing but two Cedar Mill cores sharing the same package, just as Smithfield is only two Prescotts sharing the same core. Shrinking to 65nm frees up a considerable amount of space on the die, so moving to 2MB L2 cache seems logical. Also, all 9xx processors will have EIST enabled, whereas only the 840 and 830 Pentium D processors have EIST. That makes sense for 8xx, as EIST currently just drops the CPU speed to 2.8 GHz; we would hope that the 9xx models will improve the functionality of EIST. EM64T and XD are supported on all upcoming processors. The last final addition to the 9xx line is VT, or Vanderpool Technology.

We have discussed Vanderpool in the past (first in 2003), but it seems like there is still a large amount of confusion regarding the technology. Vanderpool - and AMD's competing technology, "Pacifica" - enable a CPU to run multiple operating systems on a single CPU at the same time. The particular demos that we have seen show four operating systems running on a single machine independent of each other. Unfortunately, this requires BIOS support, Chipset support and Processor support, not to mention OS support. Intel claims that VT will roll out on the 6xx processor line before the end of the year, so we expect 945 and 955 to fully support VT as of now, since the next chipset revision from Intel won't come until Q2'06 (Broadwater).

Intel Single Core Performance Desktop Lineup LGA775

Processor

Speed

L2 Cache

FSB

Launch

Pentium 4 672

3.80GHz

2MB

800MHz

Q4'05

Pentium 4 663

3.60GHz

2MB

800MHz

Q1'06

Pentium 4 662

3.60GHz

2MB

800MHz

Q4'05

Pentium 4 653

3.40GHz

2MB

800MHz

Q1'06

Pentium 4 643

3.20GHz

2MB

800MHz

Q1'06

Pentium 4 633

3.00GHz

2MB

800MHz

Q1'06

Pentium 4 631

3.00GHz

2MB

800MHz

Q1'06

Pentium 4 571

3.80GHz

1MB

800MHz

Soon

Pentium 4 561

3.60GHz

1MB

800MHz

Soon

Pentium 4 551

3.40GHz

1MB

800MHz

Soon

Pentium 4 541

3.20GHz

1MB

800MHz

Soon

Pentium 4 531

3.00GHz

1MB

800MHz

Soon

Pentium 4 521

2.80GHz

1MB

800MHz

Soon


With 13 new performance SKUs in the next year, Intel certainly has its work cut out. The new 5x1 processors are set to launch before the end of this month, although the only real advantage that they have over the existing 5x0 and 5x0J chips is EM64T support. The interesting processors (highlighted in bold) are the upcoming 65nm Cedar Mill cores. Roadmaps hint that Cedar Mill will top out at 95W TDP per core and 3.80GHz is the highest clock on the roadmap. The Pentium 4 551 has a TDP of 84W. The roadmap also reveals that the prices of the 6xx line will cut dramatically mid-August, and VT enabled Prescott 2M (Pentium 4 672 and Pentium 4 662) will retain a premium over their non-VT counterparts.

As we mentioned earlier, two Cedar Mills running at lower clock speeds will compose a Presler dual core processor (Smithfield's replacement). HyperThreading, EIST, XD and EM64T are enabled on all of these new processors. Furthermore, all Cedar Mill chips will also receive VT technology, with the exception of the Pentium 4 631. The Pentium 4 631 and 633 are identical except the absence of VT in the 631. We're not entirely sure why the 631 is even being produced, other than to perhaps fill a niche market. All of the Cedar Mill cores clearly have VT support (as do the Prescott 2M cores), so why Intel would want to deactivate it in one model is anyone's guess.

With both 65nm Presler and 65nm Cedar Mill, the interesting thing to note is how low the FSB clock speeds remain. We had almost expected next generation Pentium processors to ramp up to 1066FSB after the most recent Pentium 4 EE utilized 1066FSB. In fact, Intel relaunched its 925X motherboard lineup (with 925XE) to support the faster bus, and all current generation motherboards make a big deal of supporting 1066FSB. As our tests revealed, 1066FSB did virtually nothing to improve performance on the 3.46EE over 800FSB. The conservative FSB speeds on Presler and Cedar Mill indicate to us that Intel probably did nothing to shorten Prescott's pipeline, although there may be other reasons why they have decided to keep the bus speed low as well.



Also mentioned in the roadmap were speed and feature revisions on the Celeron lineup. Aside from the extra speed boost, the new Celeron chips will also receive EM64T support.

Intel Single Core Value Desktop Lineup LGA775

Processor

Speed

L2 Cache

FSB

Launch

Celeron D 355

3.33GHz

256KB

533MHz

Q4'05

Celeron D 351

3.20GHz

256KB

533MHz

Soon

Celeron D 346

3.06GHz

256KB

533MHz

Soon

Celeron D 341

2.93GHz

256KB

533MHz

Soon

Celeron D 336

2.80GHz

256KB

533MHz

Soon

Celeron D 331

2.66GHz

256KB

533MHz

Soon

Celeron D 326

2.53GHz

256KB

533MHz

Soon


The Celeron D 351/350 will launch this month at $127 with price cuts on all Celeron and Pentium chips almost exactly a month after. Unfortunately, our crystal ball doesn't go past Celeron 3.33GHz. We would expect to see a Cedar Mill revision of Celeron, perhaps with 512KB L2 cache. The roadmaps very specifically do not show any new value processors based on 65nm at least through Q2'06. The roadmap does hint at speed bumps in Q2'06, but the exact reason why there are no 65nm value processors seems quite vague.

Desktop Chipsets

The roadmap also starts to talk about Intel's Broadwater chipset. Broadwater sounds exciting because it replaces all chipsets for Intel - from 955X all the way down. Our guess is that Broadwater will act more like nForce; different revisions will fill differing demands. Where Intel always used to speak of two differing chipsets (like Canterwood/Springdale, Alderwood/Grantsdale, Glenwood/Lakeport), even if they were nearly identical, the fact that Intel talks about a single chipset family unifying all of their desktop platforms indicates that things won't be exactly business as usual come Q2'06 during the next chipset launch. Aside from the general updates (ICH8 and next generation iAMT), the roadmaps revealed almost nothing about Broadwater.

Just before Broadwater, we will see the launch of 945GZ. G45GZ seems almost like a step back, with identical features to 915G including 800FSB and DDR2-533. However, the chipset will get an update on the integrated graphics to GMA950 and an updated Southbridge to ICH7. Oddly enough, Intel also claims that this will be a mainstream chipset, even though the FSB and DDR clocks are lower than existing 945P products. If anything, this might be just another indicator that Intel's push for 1066FSB wasn't really the solution that they had intended.

Intel has also decided to rework their motherboard SKUs and this should be evident already on the retail market. Each new Intel branded motherboard based on 945 or higher will receive one of several tags listed below:
  • X - Extreme Series
  • M - Media Series
  • E - Executive Series
  • C - Classic Series
Although the X and M are pretty self-explanatory, it looks like the E and C ratings seem a little ambiguous. Judging from Intel's website, it seems like there is a bit of overlap between some of these indicators. Intel's roadmap was also very pleased to announce that all of their current motherboards and future motherboards are lead free. There was also a bit of surprise that Intel will continue to work on new 915G designs right up until 945GZ. Either 915 is pretty comparable to 945 for value systems or Intel just has a lot of 915G chips left that they want to get rid of.



Yonah Yonah Yonah

It sounds like it should be part of a song, but really, it's just the core name of Intel's most promising dual and single core approaches that will launch in Q1'06. Anand gets uncomfortably giddy whenever someone mentions Yonah, although some of the revelations like clock speed were a large letdown to us. Yonah is definitely something that we talk about a lot and the 65nm dual core processor based on an evolved Dothan is really exciting. Even with the letdown on clock speed, there are more SKUs than we had originally thought, which should make low end laptops and media centers really competitive on the low end. Media centers, you say? Yes, it looks like Pentium M finally does have some sanctioning by Intel for use outside of laptops and blades. The bold chipsets indicate discrete graphics only.

Intel Single Core Value Desktop Lineup LGA775

Chipset

FSB Clock

Memory Clock

Launch

955XM

667MHz

DDR2 667MHz

Q1'06

945GM

667MHz

DDR2 667MHz

Q1'06

945PM

667MHz

DDR2 667MHz

Q1'06

945GMS

667MHz

Single Channel

DDR2-533

Q2'06

940GML

533MHz

DDR2-400

Q2'06

915GM

533MHz

DDR2 533MHz

Soon

915PM

533MHz

DDR2 533MHz

Soon

915GMS

400MHz

Single Channel

DDR2 400MHz

Soon

910GML

400MHz

DDR2 400MHz

Soon


Intel is launching two chipsets dedicated specifically for small form factor notebooks and PCs; 915GMS (soon) and 945GMS (Q2'06). 915GMS utilizes single channel DDR2-400 and 400FSB, while the much more powerful 945GMS will use single channel DDR2-533 and a 667FSB. For laptops, avoiding dual channel memory isn't a bad idea, but judging by the performance increase that we saw when running Dothan on an 865PE motherboard using ASUS' adapter, Pentium M can certainly make use of additional memory bandwidth - compression, games, and workstation tasks all showed pretty significant performance increases. We got a small taste of Pentium M in the digital home at Computex this year with some demonstrations of 915GMS from manufacturers like Shuttle and Intel. Don't expect HTPCs all over to start using Pentium M in troves, but at least it's a win for those who enjoy Pentium M over Pentium 4 and Pentium D.

Intel's integration of 945 and 955 into the next generation Centrino platform (also known as Napa) will come in three main flavors (945GM, 945PM and 955XM) with 945GMS taking up the SFF route a quarter later. Napa gets all the function from each of the existing chipsets, but also adds iAMT to the Yonah processor, Vanderpool, 3945ABG wireless and Gigabit Ethernet. This all has us very excited until we caught a glimpse of the launch speeds and prices.

Intel Dual Core Mobile Lineup LGA775

Processor

Speed

L2 Cache

FSB

Launch

Cost

Pentium M x50

2.16GHz

2MB

667MHz

Q1'06

$637

Pentium M x48 LV

1.66GHz

2MB

667MHz

Q1'06

$316

Pentium M x40

2.0GHz

2MB

667MHz

Q1'06

$423

Pentium M x38 LV

1.50GHz

2MB

667MHz

Q1'06

$284

Pentium M x30

1.83GHz

2MB

667MHz

Q1'06

$294

Pentium M x20

1.66GHz

2MB

667MHz

Q1'06

$241

Pentium M TDB

1.66GHz

2MB

667MHz

Q2'06

$209

Pentium M TDB LV

1.20GHz

2MB

533MHz

Q2'06

???

Pentium M TDB LV

1.06GHz

2MB

533MHz

Q2'06

???

Pentium M 780

2.26GHz

2MB

533MHz

Q2'05

$637

Pentium M 770

2.13GHz

2MB

533MHz

Now

$637


The new Yonah chips are denoted with an "x" in front of their product name because we do not know where they will fall into Intel's product naming yet - although 8xx or 9xx would be the best candidates. There are two surprises here, the first obviously being the low clock speed. We had expected a higher clock than the existing Pentium M chips, much in the same manner that Dothan is capable of higher speeds than the earlier Banias chips. However, just as Cedar Mill and Presler come with similar clock speeds to their 90nm predecessors, Yonah is initially slated to launch at about the same speeds as current Dothan parts. The clock ramp will surely come eventually, but don't expect phenomenal clock speeds particularly for a first generation. Intel claims that the TDP for 2.0GHz Yonah will be around 31W and 15W for the Low Voltage version.

As a dual core solution, Yonah is the most advanced (other than perhaps Itanium 2 Monticeto) solution that we have seen out of either AMD or Intel. This has a lot to do with the fact that Yonah isn't just two cores slapped together (notice that they share the same cache). It is being built from the ground up as a dual core solution, similar to how Banias was designed specifically with the goal of low power and mobility. We have high hopes that it will realize better performance scaling than some of the other Intel dual core chips. Here's where things take an interesting twist.

The second big surprise are the "TBD" (To Be Determined) chips. These are single core Yonahs. Since all the original documentation about Yonah claimed that the two cores were intertwined, our guess is just that the single core versions are identical to the dual core versions with a single core disabled. Given the added complexity of a second core, we wouldn't be surprised to find that the single core Yonahs will initially be composed of chips with one faulty core - rather than throw out the whole core, Intel can just deactivate the faulty half and sell it at a reduced price. We've been seeing this for quite a while with reduced cache versions of some processors, and it makes sense from a manufacturing and yield perspective. For $209, however, a single core 1.66GHz Yonah would have to have some pretty amazing performance increases over the existing Pentium M 740 and 735 that cost just over $200 today. It looks like we will find out a year from now.

Yonah has other endeavours as well, including a server variation on the chip (Sossaman) and Celeron M. Sossaman will begin to replace Low Voltage and Ultra Low Voltage versions of Xeon as early as Q1'06. Pentium M has already proven itself extremely valuable in the blade market, so dual core, dual processor configurations seemed almost inevitable. The first dual processor configurations of Sossaman are expected in Q2'06. Yonah already has some extremely interesting design features, but whether or not they scale to two or more processors is something that we definitely plan on exploring more in the future.

Celeron M for Yonah will have 1MB L2 cache and run at 533FSB. Unfortunately, it does not appear that Yonah Celeron M will utilize EIST. Details on Celeron M seem very sparse, but we do know that the new Celeron M lineup will start 4xx.



The Server Side of Things

With the exception of Yonah showing up where Xeon should be, there is very little new for servers.

Dual core is inevitably moving toward servers, and the chip that will spearhead that launch is Dempsey. Dempsey is similar to Smithfield, but at the same time, it offers HyperThreading support, 1066FSB, Demand Based Switching and Vanderpool Technology. Expect Dempsey to show up around Q1'06. The enterprise version of Dempsey, dubbed Paxville, will act as the enterprise large cache version of Dempsey, but at a slower 800MHz front side bus. Expect Paxville at about the same time as Dempsey. Intel makes specific note on the roadmap that the processor brand name for Paxville and Dempsey is TBD - maybe Xeon has had its end?

Intel Single Core Volume Server Lineup LGA775

Chipset

FSB Clock

Cache Size

Launch

Xeon 3.6GHz

667MHz

2MB L2

Q4'05

Xeon 3.4GHz

667MHz

2MB L2

Soon

Xeon 3.2GHz

667MHz

2MB L2

Soon

Xeon 3.0GHz

667MHz

2MB L2

Soon

Xeon 2.8GHz

667MHz

2MB L2

Q3'05


Until dual core shows up on servers, we have to settle for Iriwindale (Prescott 2M) server chipsets. Since Nocona 3.8GHz and 4.0GHz were canceled, there hasn't been much news on the server lineup. All new 3.4GHz and faster Irwindale chipsets will receive support for DBS (Demand Based Switching), but otherwise remain identical in core to their slower alternatives. Potomac and Cranford have all of their SKUs announced until Paxville comes along next year.

Dempsey and Paxville will need a new platform to run on. As in chipsets past, there are two next generation chipisets for server motherboards: Greencreek (the successor to E7535), and Blackford (the successor to E7520 and E7320). Blackford and Blackford VS are the base chipsets supporting dual core server processors. Blackford will support 4 FBD channels, 3 PCIe x8 segments and a total of 64GB of memory. Vanderpool Technology is supported on the motherboard as well as iAMT. Greencreek differs slightly by using two of its PCIe x8 segments for an x16 PCIe graphics slot and a snoop filter. Both chips also support 64bit PCI-X and PCI.

Closing Thoughts

There are clearly some interesting things moving forward inside Intel. The Sossaman project is probably one of our more favorite tidbits - four core Yonah blades would certainly pique our attention. The slightly lower clockspeed has us concerned about whether or not Yonah will really be able to compete with similar offerings from AMD and even Pentium D processors at the same time, but the incredibly low power requirements are enough to impress anyone.

The unified Broadwater family scheduled to replace 945P and 955X a year from now also has our attention. There was some speculation several months ago about Intel unifying their Xeon and Itanium socket design within the next year or so. While the roadmaps certainly don't indicate anything like that, unifying the desktop chipset families first might be a step toward that sort of unification.

Log in

Don't have an account? Sign up now