Saturday, October 28, 2006

Intel's Bluff (Q3 06 Results)

It's taken me a little time to go over Intel's finances in detail. However, far from being the glowing outlook that nearly everyone is saying, Intel's financial position is much more precarious. The good third quarter results were actually only a temporary fix created by slashing stock buyback. Unfortunately, even following the buyback patterns from two years ago, this will still leave Intel about $2.5 Billion short at the end of the year. The best that Intel can hope for is that it has a much better fourth quarter than even their best projections.

The overly positive outlook for Intel seems to be based on three assumptions:

  1. Intel is regaining share.
  2. Intel is making more money.
  3. Intel is back to playing its own game.

I've seen some very poor attempts to give Intel credit for a supposed increase in volume share by using a bit of statistical sleight of hand. The statistics for the last three quarters for microprocessor volume are:

Intel - 74.3%, 72.9%, 76%
AMD - 21.1%, 21.6%, 23%
VIA - 4.6%, 5.5%, 1%

Based on these percentages, the naiive claim is that Intel gained 3.1% in the third quarter while AMD only gained 1.4%. This would seem to suggest that Intel gained twice as fast as AMD in the third quarter and, presumably, this would be an indication that Intel is taking back share. Unfortuately, this distorted view is caused by the interference from VIA. VIA only sells at the very low end of the X86 market. Adding their share in with AMD's and Intel's is a bit like adding bicycles to motorcycle sales. We can get a clearer picture by removing VIA from the volume statistics and just showing the volume share between AMD and Intel:

Intel - 77.9%, 77.1%, 76.8%
AMD - 22.1%, 22.9%, 23.2%

The undistorted statistics tell a very different story. We can see that Intel has been slowly losing share to AMD for the last three quarters; quarter three was no exception. Microprocessor revenue share is similar:

Intel - 82.8%, 81.9%, 81.7%
AMD - 17.2%, 18.1%, 18.3%

Clearly, Intel is not regaining share in either microprocessor revenue or volume. Similar distorted claims are based on operating income and cash:

Q2 - Revenue - 8009, Op. Inc. - 1072, 13.4%
Q3 - Revenue - 8739, Op. Inc. - 1374, 15.7%

The operating income as a percentage of revenue appears to be up and Intel boosted its cash by nearly $700 Million in the third quarter. This looks very positive for Intel. However, there is a problem. People who buy Intel stock rely on Intel's stock buybacks to increase demand. In 2004 Intel spent $7.5 Billion on stock buyback and spent $10.6 Billion in 2005. Based on the lower 2004 pattern Intel should spend $7.7 Billion this year. However, Intel has only spent $4.4 Billion so far and should have spent about $1.9 Billion in the third quarter. Instead, Intel only spent $500 Million.

It is clear then that Intel boosted its third quarter numbers by slashing its stock buyback. This makes the third quarter numbers look good but puts Intel $1.4 Billion behind in addition to the $1.9 Billion it would need to spend in the fourth quarter. However, if Intel spends the same as it did this quarter it will end up $2.5 Billion short which will cause its stock to plunge. If it pulls the cash back out that it just put in this would end up only $1.2 Billion short. This is a tough position for Intel. If it doesn't spend any cash then its stock buyback will drop below the 2004 levels and will only be a little ahead of the 2003 levels. This would be a 50% reduction from the 2005 levels and something the stock market surely would not overlook. This could easily cause Intel's stock to slump when the end of year results are released(which will be sharply down from last year). Unfortunately, Intel will not have the money to spend so it will have to choose between spending cash and risking a drop in stock price. From this we can see that Intel's money position is currently worse than it was even in 2004. Intel has not yet recovered in terms of money.

The final assumption is that Intel had to play AMD's game while AMD was ahead in processor performance and now that Core 2 Duo is ahead Intel can play its own game. Well, let's assume for a moment that this is true, that C2D's superior performance is creating lots of demand for Intel and eroding AMD's demand. This could cause AMD to drop its prices even lower to avoid losing too much volume share and this would hurt AMD's margins even more.

The problem is that AMD hasn't lost in volume share; Intel actually lost a little. If the assumption about demand were true then this wouldn't fit with the facts at all.We know that AMD was capacity limited in the third quarter. If demand were strong then Intel should have had no trouble picking up customers beyond what AMD was able to supply. But, this didn't happen. Secondly, the fact that Intel's margins were lower than AMD's and will stay lower in the fourth quarter disproves the notion that AMD has to compete by dropping prices. A similar idea was that Intel was going to get rid of its overstock by selling it at reduced prices and this was supposed to undercut AMD and cause further loss of margin. However, AMD's margins didn't drop all that much and Intel ended up having to take a loss on $100 Million worth of overstock. Apparently, this was only part of the overstock and Intel may have additional losses in the fourth quarter.

The fact that Intel was unable to gain volume share while AMD was capacity limited disproves the idea of extra demand for Intel chips. The fact that Intel didn't gain at all in terms of revenue share also disproves the idea that Intel's chips are more valued. Intel finds itself faced with a competitor who is serious about growing capacity. The only way of stopping this would be to somehow reduce demand for AMD chips and this hasn't happened. Intel seems to be overlooking the fact that having two FABs plus second sourcing from Chartered makes AMD a very reliable supplier; Intel has lost this advantage. Intel has been the strongest player by far in terms of integrated graphics for both desktops and mobile. However, now that AMD has ATI Intel faces a whole new level of competition in these markets.

AMD is currently taking mobile share from Intel. Many have assumed that Merom, the mobile version of Conroe would easily hold this market for Intel. However the fact that Intel will keep making the lower performing Yonah until the end of 2007 suggests otherwise. It appears that Merom is having more difficulty with power consumption than the older Core Duo. AMD's Turion should compete with this very well as it moves to 65nm. This seems to have been a strategic mistake for Intel and not one that it can fix soon. Faced with an inability to manage multiple processor projects, Intel lumped everything together with Core 2 Duo. This chip tries to do the job of server, desktop, and mobile. Apparently, it isn't doing mobile as well as it was expected. This unfortunately occurred at the same time that AMD decided to split mobile off into a separate family. This seems to be what will give AMD the ability to take share in mobile away from Intel. Intel surely realizes its mistake and is probably frantically trying to reform another all mobile team as it had with Pentium M. However, this is likely to set Intel back into 2008 and it can lose a lot of share during that time.

The one size fits all strategy with C2D is also not working as well as it could with servers. Woodcrest doesn't do 4-way and Intel won't have a 4-way until Q3 07. This means that during this time Intel has to compete with the outdated Cedar Mill based Xeons. The result has been that Intel has not been able to take back any of the 4-way and higher share that it lost to AMD both from Xeon and Itanium. Given AMD's recent wins in supercomputers with Opteron it seems that AMD will remain strong in the top end of servers well into 2009. Although Intel has made some gains in 2-way servers it will soon be faced with both lower power 65nm based Opterons and a native quad core design with K8L. This is a fairly grim prospect for Intel as it won't give them much time to gain any share and even then only at the bottom. Further, Intel appears to be behind in base architecture and it doesn't appear that they will catch AMD until 2009. AMD has a good chance of reversing any server losses during that time and taking additional server share.

The sole bright spot for Intel is the desktop. Or maybe it isn't. This depends on how much of Intel's third quarter volume came from moving lower priced P4 stock. It could be the case that Intel's rise in desktop volume share was only a one quarter bump and will drop again in the fourth quarter or even the first quarter of next year after it moves more overstock in the fourth quarter. At this point though there isn't any way to tell. It seems reasonable though that, with increasing pressure from AMD within the corporate market with its Stable Image Platform and integrated graphics from ATI, AMD is not likely to lose very much share. Its deals with the two highest retailers in China and its deal with Dell also suggest that any volume loss on the desktop will be limited. Some have tried to argue that Dell is rapidly losing ground and therefore won't help AMD. However, this seems to be more of a sour grapes argument than anything else. Looking over Dell's financials I was unable to see more than a drop of about 4% and this is hardly proof of any downhill slide for Dell. Intel dropped much more than this during 2006.

Although it has been suggested that AMD's current capacity constraints will cause it to lose customers this logic too does not appear very sound. AMD's capacity will increase rapidly next year so Intel can't count on capacity constraints for AMD unless the market grows much more rapidly than it has. Then too, Intel still would have to turn these limitations into demand for its own products as it was unable to do in the third quarter.

Intel is also unlikely to improve its popularity with vendors. Even though AMD now has ATI it is quite willing to let its customers use other brands of chipsets if they prefer. Intel in contrast, is much more insistent about pushing its own products as it has with its exclusive chipset policy for Centrino.

I have been unable to find any evidence in the third quarter reports that Intel's prospects are better than AMD's. Intel has been unable to create demand or decrease demand for AMD. Intel will find itself in a defensive position trying not to lose share in mobile and corporate accounts while trying to push a single chip design that is not fully adequate for either server or mobile use. It has to do this against a tough competitor who is putting huge amounts of its earnings into capital expenditures to grow its capacity. On top of this, Intel will find itself with a very disappointing 2006 earnings report and enormous pressure on stock buyback with money that it won't have. It is likely to be faced with the prospect of either spending down its cash even more or watching its stock price plunge. From all of these disadvantages it is clear that Intel is not playing its own game.

Sunday, October 22, 2006

Tigerton or Kittenton? Memory Amnesia.

The glowing reports about Intel's roadmap for 2007 continue. Central in this enthusiasm is the native quad core Tigerton which is supposed to be the platform that will keep Intel ahead of AMD's K8L technology. However, in evaluating this platform most people seem to have forgotton the memory bandwidth lessons from P4 and Athlon 64. This amnesia of memory requirements has led to exagerated expectations of performance for Tigerton.

The diagram above for Tigerton looks impressive. Surely four buses with an aggregate bandwith of 34 GB/sec would be enough. Or would it? Just think back to 2002 and 2003 when P4 was so strong against K7. I remember how Intel enthusiasts bragged about the 800Mhz FSB for P4 which was well ahead of what K7 had. Curiously, while these same Intel enthusiasts thought 800Mhz was necesary for one core they now seem to be insisting that 1066Mhz is enough for four cores. Let's consider this more carefully.

We know that when Intel increased its FSB from 400Mhz to 800Mhz that it didn't use all of the bandwidth. We also know that Athlon 64 seemed to get along okay on socket 754 which only contained one memory channel. One memory channel on socket 754 was basically the same as the 400Mhz FSB for P4. It is difficult to imagine that anyone today would suggest that this was more bandwidth than these processors needed. I think clearly that this level for one core is the most reasonable starting point. Let's see where this takes us.

If we need one memory channel with DDR 400 for one Athon 64 core then the dual memory channels with socket 939 and 940 would seem to be enough for X2. Likewise, an 800Mhz FSB with two channels of DDR 400 or DDR2 400 should be enough for Intel's dual core Preslers, Conroes, and Woodcrests. If this is the standard then both AMD and Intel clearly have more bandwidth than necessary for dual cores. AMD's Integrated Memory Controller (IMC) allows expansion beyond two sockets and Intel's 5000 series chipset with two buses again allows expansion to two sockets with similar memory bandwidth. Dual core today is well in hand. Let's look at more cores.

Intel will soon release the quad core Kentsfield and AMD will release the quad core Barcelona in Q2 07. The above diagram is for Tigerton which most likely would arrive in Q4 07 but could arrive in Q3 07. Although Intel does say Q3 07 the chips may not arrive in volume for another two months if recent releases are any indication. At any rate, we know that dual core works well but will quad core work the same? If an 800Mhz FSB is required for dual core then reasonably a 1600Mhz FSB would be needed for Kentsfield or Tigerton. We know that Kentsfield will have a 1333Mhz FSB at most and may only have 1066Mhz. A FSB of 1333Mhz would only be equivalent to a Conroe with a 667Mhz FSB. This would be equivalent to a 333Mhz FSB for P4. We can probably skip comparing with a 333Mhz FSB single core P4 because it is possible that the longer pipeline changed the bandwidth requirements. Since Core 2 Duo (C2D) is not available in single core it seems reasonable to compare with the single cored Athlon 64 which also has shorter pipeline. For Athlon 64 this would be the same as using the single channel socket 754 and using DDR 333.

It does seem curious that no one uses a Conroe with the FSB set at 667Mhz since this is what is being claimed is adequate for Kentsfield. However, the question of whether or not Kentsfield can do 1333Mhz on its FSB is moot since the above diagram for Tigeron only shows an aggregate of 34 GB/sec for four buses. 34 GB/sec for four buses is only 1066Mhz per bus. This is only equivalent to a Conroe with a 533Mhz FSB or a P4 with a 266Mhz FSB. The equivalent for AMD would be an X2 or Athlon 64 with DDR 266 on socket 754. To make things worse, the memory speed listed in the diagram is incorrect. To keep up with four FSB's running at 1066Mhz, Intel will need DDR2 (or DDR3) running at 1066Mhz as well. The only way that the above diagram would work with 667 Mhz memory would be if the number of memory buses is wrong and it actually has six memory buses instead. However, even if the memory is able to keep up the FSB speed is inadequate for four sockets.

As far as I can tell the only reason for this recent outbreak of amnesia for memory requirements is the sloppy testing over at Tom's Hardware Guide. Apparently, THG is on some kind of crusade to prove that Intel has the best hardware even if THG has to fake the tests to prove it. It doesn't seem to do any good to criticize the testing at THG because the excuses from Intel enthusiasts seem to grow exponentially in proportion to how bad the testing is. However, not matter how sloppy or unprofessional or biased the testing is at THG there really is no escaping common sense. A Tigerton with four buses of 34 GB/sec aggregate bandwidth is only equivalent to a Conroe with a 533Mhz FSB. I'm curious why THG doesn't run a Woodcrest test with the FSB set to 533Mhz to show what the expected performance with Tigerton will be. Tigerton should be double a Woodcrest with 533Mhz FSB. The comparative proof for Kentsfield is easy as well since you only need to set the FSB to half the FSB speed for Kentsfield to get the same memory bandwidth.

The notion that a Tigerton can run four cores with a FSB of 1066Mhz is beyond ridiculous. This would only be 533Mhz for Conroe and equivalent to DDR 266 for both X2 and Athlon 64 on socket 754. Even the old Barton Athlon XP 3200+ used a FSB of 400. To get as low as Tigerton we would have to go all the way back to the Athlon 2600+ which used a 266Mhz FSB. None of the P4's were ever this low as even the original 1.4Ghz Williamette's had 400Mhz FSB's. We would have to go back to the PIII to get down to 266Mhz on the FSB.

It is clear that memory bandwidth is going to be a serious problem for Intel and is going to set Tigerton's performance back badly. The question then is how well AMD stacks up in the memory bandwidth race. The socket 753 Athlon 64 used one channel of DDR400 memory. This was doubled to two channels with socket 939 and 940 so this would be the same equivalent bandwidth for X2. Presumably, Barcelona would need double this or 800Mhz memory. This shouldn't be a problem as even the memory controller on Revision F is about to handle DDR2 800. This would be equivalent to 1600Mhz and well ahead of the 1066Mhz for Tigerton. In fact, a little math shows that this is fully 50% more bandwith than Tigerton will have. This means that in memory intensive operations Tigerton will bog down to 33% slower than Barcelona or you could also say that Barcelona will be 50% faster.

Looking at the Caneland chipset I can say that moving to quad buses is a step up from the limited dual bus 5000 chipset. I can also say that the 64 MB snoop filter cache is a great idea. This is basically a map that prevents cache coherency traffic from being retransmitted across to other buses unnecessarily. However, this is bad because it indicates that the FSB will still carry the cache coherency messages. The snoop filter won't help when the cache coherency messages must be transmitted to another bus. It is not nearly as good as AMD's design where none of the cache coherency messages effect memory bandwidth. It is also disappointing to see that Intel will still be using the outdated MESI protocol rather than a more advanced protocol like MOESI as AMD uses. This too would reduce cache coherency saves and loads. When another chip needs to access a memory location that is in another chip's cache, MESI requires that chip that has the data save it to memory whereas MOESI can allow sharing without having to save.

The Tigerton server chip is an upgrade to Woodcrest and will at least give Intel something to replace its aging Cedar Mill Xeon chips and finally let C2D go beyond 2-way. However, because of memory bandwidth constraints, Tigerton will be proportionately slower than Woodcrest. Secondly, the octopus-like design of the Caneland Northbridge with four FSB's and four memory buses is not going to be compact or inexpensive. Finally, this limited four bus design will not allow Tigerton to move up to bigger 8-way and 16-way server designs as AMD will have. In conclusion, it appears that although Tigerton looks good, it will be held back by the memory bandwidth of the Caneland platform which is as low as what was used by PIII and Athlon 2600+. This will cause Tigerton to get beaten badly in memory intensive applications, making it more of a kitten than a tiger. Clearly, memory bandwidth will be Intel's biggest problem into 2009.

Thursday, October 19, 2006

A Real Forecast

Although there have been many projections lately that favor Intel, it is my guess that these are based largely on old biases towards Intel. I believe that a more careful look at AMD and Intel shows that AMD's performance and current position are better than Intel's. And, although I believe that Intel hit bottom in the last quarter and will now grow, it seems likely that AMD will grow faster during 2007 and 2008.

For recent performance we can see that AMD clearly hit its stride in the fourth quarter of 2005. This quarter was the best quarter in Intel's entire history. Yet, AMD was able to increase its processor revenue share from 12% to 15% during this same quarter. It is staggering to think that AMD pulled revenue away from Intel while Intel was at its best. However, these gains by AMD left Intel unable to rid itself of its inventory overstock which increased 11% from the previous quarter. In contrast, AMD reduced its own inventory by 58%. Intel began to drop and hit bottom just two quarters later in Q2 06. Intel has been unable to rid itself of its overstock and it has continued to build since Q3 05. Today, AMD's inventories are less than half of what they were in Q3 05 while Intel's are 60% higher.

The best that can be said for Intel is that they've hit bottom and are slowly growing again. However, the frantic scratching in the dirt by some in an attempt to put a positive spin on Intel's numbers is truly amazing. I've seen people suggest that Intel is now doing better than AMD. Unfortunately, statements like this are purely delusional. The reality is that 2006 is a very bad year for Intel. This will be the largest processor revenue drop for Intel since 2001. Short of a miracle (which even Intel does not project) in Q4 06, Intel's total yearly processor revenues for 2006 will be much lower than 2005's and only about equal to what Intel received in 2004. In contrast, AMD's processor revenues for 2006 will be double what they were in 2004. I'm baffled how zero growth in two years versus 100% growth could be equal or ahead for Intel. Clearly, AMD has done very well and Intel has not.

AMD's processor revenues are up 33% from the Q3 05 while Intel's are down 21%. Faced with dismal numbers like these, those who favor Intel have resorted to looking only at this quarter versus last quarter. However, even here there is no place to hide. Intel's processor revenue's increased 8.8% from last quarter. However, AMD's increased 10.1%. The fact that AMD's margin dropped more from Q2 06 to this quarter has been tossed out with glee by those seeking to fabricate a better outlook for Intel. Howevever, the only reason that this is true is because Intel's had already dropped badly during Q2 06. Today, Intel's margin at 49.1% is still below AMD's at 51.4%. I find it amazing that some don't see how noteworthy this is. Intel has been consistently ahead of AMD and usually be several percentage points. Further, the overall drop has been far worse for Intel. Intel's margin dropped from a recent high of 61.8% in Q4 05 while AMD's dropped from a recent high of 58.5% in Q1 06. This is a 21% drop for Intel versus a 12% drop for AMD. None of these numbers favor Intel. Yet, one doesn't have to look very far to find people proclaiming that these are great numbers for Intel and that they show a glowing future. However, there is no reason to believe the phoney outlooks and no reason to look at Intel with rose colored glasses.

AMD has had very strong demand (up 50%) for its mobile chips. Unfortunately for Intel this is not likely to improve. AMD's 65nm mobile chips will be out soon and will have lower power draw which is a good thing for mobile. AMD is also certain to get a boost in mobile chipsets from ATI which takes away the strongest advantage that Intel has enjoyed with its Centrino line. A loss of share in notebooks is something new as this has been the area that Intel has held onto the strongest. Similarly, AMD's very aggressive move into integrated chipsets gives AMD a good boost for both corporate client computers and budget systems. The budget category lines up well with AMD's recent partnership with the two largest computer retailers in China. The client systems are likely to be a good complement to the already increased offerings by HP, Dell, Sun, and IBM in AMD servers.

Intel is pressuring AMD hard in 2-way servers. This is not surprising as Woodcrest has impressive SSE performance and works quite well with dual core and dual socket with the dual bus 5000 chipset. The only bad part is that this chipset uses FBDIMM which already has a limited lifespan. Worse still, this technology will not grow large enough to become profitable while Intel will have to continue support to the many customers who have already bought or are buying these systems. This will be a future drag on Intel although it only a small one. AMD would pull ahead on 2-way systems in Q2 07 except Intel should have a non-FBDIMM chipset available which will remove the disadvantages of the FBDIMM technology. So, AMD should be about equal on 2-way dual core systems which will prevent Intel's gaining back share. However, Intel will fall behind on quad core systems and will not be competitive on 4-way sooner than 2009. This is the case because, without CSI, Intel cannot use an integrated memory controller. Without both of these Intel simply does not have the base architecture to handle dual core 4-way or quad core 2-way or higher nor has Intel suggested upgrading to a quad bus chipset. AMD should continue to make gains in 4-way and should start making gains in 2-way after Q2 07. This does at least give Intel time for a couple of good quarters with 2-way however it is unlikely to make any strong reversals in servers. Worse still, Intel then loses the advantage until sometime in 2009.

The desktop is Intel's strongest position in the near term. Kentsfield will have little effect for Intel since this is a very small market and will be countered at the top end by 4X4. Secondly, Kentsfield will be countered in Q3 07 when AMD releases the native Altair quad core. Judging from the roadmap for Pentium E 1000 and Celeron D 400 chips, the desktop clock speeds will not increase as quickly as some have suggested. AMD should be able to counter these with Rev F and G chips while competing at the high end with dual core versions of K8L. There does not appear to be any advantage for Intel in Q3 07. This would suggest that Intel's sole strategy for pulling back ahead of AMD will be a move to 45nm. However, this move is not likely to be as successful as some have suggested. In 2008, less than half of Intel's FABs will be 45nm while both of AMD's 300mm FABs will be ramping to 45nm. The advantage of APM will finally be felt as this will make extremely rapid transfers between the two FABs possible. Intel on the other hand will still require two months to move a chip or process from one FAB to another; AMD will be able to do this the same day. Two similar FABs should allow AMD even tighter control over costs although AMD's costs will fall steadily through 2007 as FAB 36 continues to ramp. The costs for 65nm on the 300mm FAB 36 are far lower than the costs at 90nm on the 200mm FAB 30. Unfortunately Intel will not see this same reduction in cost as it is already using 300mm FABs and will already have converted most of its capacity to 65nm by year's end. These reductions in cost should pull up AMD's margin even if Intel continues the price war.

I believe that many (including people at Intel itself) still have not come to grips with the reality that Intel can no longer grow its way out of trouble. Further, Intel is facing a competitor who now has two FABs and is using 300mm. In 2002 or even into 2003, Intel's current price drops would have been far more effective. Today, they have merely slowed AMD's growth while hurting Intel more. Eventually, the reality that AMD now has products spread among desktop, server, and mobile as well as second sourcing from Chartered and soon chipsets from ATI will be unavoidable. It will also become clear at some point that AMD's HyperTransport consortium is paying off with a low pricing model along with the Torrenza initiative. Intel's copycat strategy is not likley to have the same benefit and will take two years to not catch up to where AMD is now. The proposed Geneseo and CSI standards are about as fast but still less flexible than the current HT 3.0 standard. It is entirely possible that HT could be at a 4.0 or even 5.0 standard by the time CSI makes it to hardware. AMD has at least two other possibilities to compete with Intel's 45nm in 2008. These include the possibility of using TTRAM or Z-RAM for cache. This is not an option for Intel as it requires SOI. Another possibility is using the GPU as a parallel coprocessor. This could negate any boost in SSE speed. Again, it is unfortunate for Intel that using a GPU for procesing is easier with HT than with using the FSB. And, whatever effort Intel does put into this would be wasted when CSI is released in 2009.

I don't see any great gains for Intel in the fourth quarter. I'm guessing AMD will increase at about the same rate and hold onto its roughly 22% volume share. I expect Intel to grow during 2007 and 2008 but I expect AMD to grow faster and increase share. I would expect Q1 07 to be Intel's high point with servers and Q2 07 to be Intel's high point on the desktop. Overall, I would expect Intel's revenue's in 2007 to match the 2005 revenues. I don't honest see how Intel could realistically take back share from AMD given all of AMD's proejcted advantages with cost, volume, and the partners and customers it has lined up. I expect AMD's margins to rise regardless of Intel's pricing strategy due to AMD's conintued drops in costs. AMD has clearly divested of the most unprofitable sections of the company while Intel has made a less serious effort. Intel may have to revisit the issue of divestment if it cannot get its margins up. I don't see any possible way for Intel to grow at 30% while AMD grows at only 15% during 2007. Intel show grow but AMD should grow faster.

These are comments I made in December 2004.

I've been thinking lately about what is going to happen in the next year. I know that some are desperately hoping and that others are at least suggesting that Intel will make a comeback. But, so far, I can't see it.

The comeback for Intel didn't happen. AMD retained its share during 2005.

It is clear that Intel's dual core strategy is threefold:

1. Release dual core Prescott chips to slow down the erosion of the workstation/low end server market. This won't actually work very well since these chips will be starved for bus bandwidth. Expect Intel to pump up the cache. This strategy would put another company out of business but with all of that spare fab capacity Intel will only take an embarassing margin hit.

This was pretty accurate. Intel released Smithfield and then followed up with Presler. The dies are large and Intel still uses the large cache approach on the Xeon units. These processors have been ineffective against Opteron.

2. Release dual core Dothan based chips to slow down the erosion of the desktop market. Intel is hoping that just one of the benchmarks will favor this chip. This is not a great strategy but it really is all Intel has since, without being able to keep up with Prescott, Intel's only other choice is to fall farther and farther behind.

Yonah based on Dothan was next. Yonah was also not very impressive on the desktop.

3. Work on something more reasonable for 2006. This is Intel's real strategy although they won't actually admit this to their shareholders, employees, or customers. Intel is praying that they can get something out the door in 2006 that will give them a fighting chance.

At the time, no one else was talking about a new design for 2006. The talk was that Intel was going to release something competitive in 2005 which did not happen.

In all honesty, I don't think Intel has hit bottom yet nor do I think AMD has hit its full stride. I think things will be better for AMD and worse for Intel in 2005. Maybe in 6 months things will look better for Intel in 2006 but right now not even that looks promising.

Since my past comments seemed to have been accurate perhaps my latest ones will be as well. It appears to me that AMD will exit 2006 in the strongest position it has ever been against Intel. This should be a good reason for a positive forecast for AMD.

Tuesday, October 17, 2006

The Phoney Forecast Game

I've noticed an interesting trend in forecasts for Intel and AMD. No matter how Intel performs the forecasts tend to be positive and no matter how AMD performs the forecasts tend to be negative. AMD has come a long way since it was a full generation behind Intel with the K5. AMD has now caught up to Intel in servers and mobile and in 300mm FAB production. AMD's current position is much more positive than it was five years ago. However, you wouldn't tend to know this from the many biased forecasts that lean toward Intel.

The main difference between a phoney forecast and a genuine forecast is that a genuine forecast is based on a reasonable evaluation of where each company is and is likely to be. These types of forward looking evaluations seek to dismiss things that are not relevant or have little effect. These evaluations can sometimes be counterintuitive because they can take into consideration activity in markets like China which are not as obvious as what is on a shelf in a computer store or what commercials are seen on tv. This is a hazy subject of orders that may have been made months before and decisions made based on the strength of partnerships and profit margins rather than benchmark scores. These types of outlooks are often accurate but dull and sometimes difficult to understand without a background in business.

In contrast, the phoney forecast is often based on a trivial surface view of the market rather than anything in-depth. These forecasts are often grounded in emotion rather than analysis and are never counterintuitive to the author since they are often based on the author's biases more than anything else. The author may be very passionate about this view and feel very deeply that this view states what the market should do (or at least what he wants it to do). These phoney forecasts have been popping up all over the place lately and typically masquerade as genuiune forecasts. The difference can be difficult to tell but one strong clue is when the facts change but the forecast doesn't. In other words, an argument will simply flip flop in order to keep the forecast in line with the current circumstances. When this happens it is a sure sign of a phoney forecast.

For example, Intel supporters were often quite vocal in stating that Intel couldn't lose ground to AMD because Dell was Intel only and Dell had a very high volume. The association of Dell with Intel gave Intel a guarantee of a high volume customer and therefore lots of sales. Curiously, these same people are now saying that AMD is being hurt because of the supposedly low prices it is getting from Dell for its chips. This is an odd argument because these same people were never concerned about the wholesale price to Dell before. This is a good example of a Phoney Outlook. What was a strength for Intel is now flip flopped to become a phoney detriment for AMD. It is simply not reasonable to assume that AMD is hurting its profits by giving away chips to Dell at fire sale prices. This is a particularly odd argument because it is often accompanied by the idea that Intel will sell its massive overstock of Prescott based chips at low, low prices. Supposedly, this will take market share away from AMD. So, what we have is the argument that if AMD is selling at low wholesale prices to Dell that this is bad for AMD and the simultaneous argument that if Intel is selling at low wholesale prices that this is good for Intel. It should be obvious to everyone that there is no real logic in these two arguments.

Similarly, when many Intel fans saw the benchmarks for Core 2 Duo, they proclaimed that this would obviously reduce the demand for AMD processors. This half baked argument of course overlooked the fact that C2D would only reach about 30% of Intel's total production by the end of the year. People making this argument could never explain how the other 70% of Intel's production was going to reduce demand for AMD chips. Nevertheless, this statement has been repeated with great confidence by Intel fans since the first preview of C2D in March. The reality today is that AMD is actually short of chips because its processors are in such high demand that it cannot suppply fill all of the orders. This fact also negates the previous assumption that lower prices for Intel's overstock would also reduce demand for AMD chips. This simply is not happening. However, the current argument is that this is bad for AMD because failure to meet all of the demand will drive customers away. So, the argument has flipped from low demand will be bad for AMD to high demand is bad for AMD. This of course makes no sense. However, it also overlooks the fact that C2D itself has been in short supply and will be in short supply probably until end of Q1 07. It is curious that it is never stated that this short supply of C2D will drive customers away from Intel.

It was frequently suggested (and rightly so) that one of Intel's strengths was its chipsets. The low power chipsets for the Centrino line have given Intel based notebooks much better battery life than Turion notebooks. It was also stated that Intel's chipsets gave them an advantage with certain customers which was also true. Now that AMD is buying ATI it would appear that these would be advantages for AMD as well. However, now the argument has shifted to suggesting that AMD will be hurt by its purchase of ATI supposedly because of the cost. We might at first try to give this argument some credit. However, as Intel has spent down its cash reserves to less than half of what they had at the beginning of 2005 there hasn't been any criticism. Generally, this type of spending of cash is viewed as irresponsible however it tends to get brushed aside in Intel's case with suggestions that it is nothing compared to Intel's revenues. However, it is similarly considered reasonable to borrow money for capital investment as long as the amount is within the expected revenues, which it is for AMD. I would have to agree that while Intel's spending of cash is not great it does fit within their projected revenues and shouldn't hurt. Similarly, AMD's purchase, while large also fits within the projected revenues. I'm guessing that the main result of this will be the delay of the New York FAB. However, the ATI purchase does seem to be more strategic than just building a third FAB and AMD's capacity will increase rapidly anyway in 2008 with FAB 38. The New York FAB could be started late in 2007 and still be finished in time to take up the slack after FAB 38 tops out in 2009. So, it seems to be the case that the argument that AMD will be hurt by the purchase of ATI is just another flip flop of logic to turn a similar strength into a phoney detriment.

The issue of layoffs has been similar. I view layoffs and divestment as a good thing for Intel and something that makes it more competitive. I think AMD's divestment of Spansion was a very good thing and something that Intel should do as well. However, lately we've seen a much different treatment of layoffs between Intel and AMD. As Intel has announced layoffs this has been taken in stride even though most expect that more layoffs will follow. There has been little criticism of Intel and I think that has been correct. However, when rumors popped up about layoffs at AMD this was pounced on as though it were some glaring harbinger of doom for AMD. This is very odd since AMD stated that there would be some layoffs when it merged with ATI because of redundancies. So, basically, unnannounced layoffs by Intel have been no big deal while expected layoffs by AMD are seen as tragic. Something is wrong with this picture.

There are similar phoney arguments about partnerships and markets. For example, it seems to be the subject of great bragging rights when Intel suggests that it has signed X number of companies for initiative Y. These were often mentioned to proclaim Intel's support and standing in the market. However, Intel has driven partners away and seen increasing delays with CSI while AMD has gained traction because of HyperTransport. So, some have shifted this argument to the notion that Intel gains by having fewer partners because it can keep more of the market to itself. In other words, in a classic flip flop it is argued that having many partners means that Intel will sell lots of chips and therefore make lots of money, however it is also argued that having fewer partners and proprietary standards means that Intel will still make lots of money. We see an almost identical argument with licensing. AMD has had no licensing and therefore low cost to use HyperTransport technology. This has been countered by the notion that this is an advantage for Intel because it makes money off of licensing. However, now that it is suggested that Intel may be reducing its licensing costs to try to compete with AMD this is now stated as an advantage for Intel. So, apparently, both high and low licensing fees help Intel.

Growth has been another area of contradiction. If AMD does not grow quickly as happened from 2002 - 2004 this as seen as an indication that AMD cannot challenge Intel. However, AMD's growth in the past two years has been staggering. AMD has doubled in size in the last two years. This is a pace that has exceeded every other top 50 chip producing company in the industry. So, unable to criticize AMD based on growth, the forecasts now concentrate on size and suggest that AMD cannot continue to grow because it is already large. In other words, if AMD doesn't grow it is bad but if AMD does grow it is still bad. In a similar vein some forecasters try hard to associate the best FAB technology with Intel. However, there are two problems with this idea. The first is that AMD's FAB 30 has been the world's number one rated FAB for five years running which is why Intel keeps its FAB ratings secret. The second is that AMD has the very sophisticated APM control system for its FABs which will be a tremendous benefit once it has two 300mm FABs. This will allow production to move from FAB to FAB much more easily than it will at Intel. So, again, unable to associate the best FAB tech with Intel the forecast shifts to capacity and it is insisted that Intel will stay in the lead simply because it has more FAB capacity. Unfortunately, Intel has cut back on its capital spending severely while AMD has increased its by 50%. This would certainly suggest that AMD inteneds to have more FAB capacity in the future while Intel is cutting back.

Those are not the only contradictory gems that have been floating around. There are many others. For example:

If ATI can't sell chipsets for Intel processors then this will hurt AMD because of the loss of Revenue to ATI : If ATI sells chipsets for Intel processors then this will hurt AMD because ATI will be helping Intel.

If AMD can't gain share in the Chinese market this will hurt AMD because this is the fastest growing market : If AMD gains share in the Chinese market this will hurt AMD because this is a low margin market.

If AMD doesn't produce quad core processors soon this will hurt AMD because Kentsfield will be out in November : If AMD starts making quad core processors this will hurt AMD because it will use up too much of AMD's capacity.

If AMD cannot increase its Average Selling Price this will hurt AMD's margins : If AMD increases its Average Selling Price this will hurt AMD's volumes.

It does seem unfair that all of the phoney arguments that I've given are in favor of Intel and against AMD. There is certainly no shortage of biased statements made by AMD fans in favor of AMD and against Intel. However, these statments that favor AMD don't tend to travel very far. It is rare to see something of this nature picked up by the general web press or projected into the forecast of a financial or stock analyst. In contrast, phoney arguments that favor Intel seem to have much more stamina and can be repeated almost anywhere and by anyone on the web including by professionals. Since I don't know of any way to stop the internet press from repeating phoney arguments and outlooks the best thing is to recognize them for what they are and treat them accordingly.

To help illustrate, I'll give an example from Information Week.

AMD Seen Strong This Year, Intel Next

This article with quotes from Merrill Lynch, Citigroup, and Reuters at first appears to be a professional outlook article. However, a closer look reveals that it is in fact phoney. The actual content of the article from Reuters is entirely biased in favor of Intel and simply seeks a rationalization to say that Intel will come out ahead. In other words, the outlook is decided before looking at the evidence and the arguments are simply used to try to fit the preferred outlook into the reality of the data. Essentially, the arguments become an adapter that allows the author to fit the square peg of Intel optimism into a round hole.

The first thing we are told is that AMD's stock has fallen because of fears that Intel's steep price cuts would reverse AMD's gains. This statement is irrelevant. AMD's performance is not directly rated to its stock value. This statement seems to work as a bandwagon fallacy to try to suggest that the author's view is actually common. The author has to use this irrelevant line because what follows does not support his bias.

Price performance continues to be better at AMD, even with all of Intel's pricing cuts.

AMD has been rapidly increasing its share of the PC processor market

Citigroup analyst Glen Yeung said he expected AMD to gain more share in 2007, thanks in part to a new supply agreement with the No. 1 PC maker Dell Inc.

"We expect momentum to favor AMD," Yeung said in a research note. "While the shares have already had a significant move so far in the second half of the year, we believe additional upside is likely through year-end."

At this point, the author has been essentially run over by the statements favorable to AMD. To try to get back to his original favored view of Intel he again talks about stock. He mentions that AMD's stock value is about the same as Intel's when compared with profit. This is statement is not only irrelevant, it is completely ridiculous. Intel moves most of its profits via stock buyback whereas AMD does not. If we were to actually follow his suggestion that the value of the stock somehow indicates a company's future performance then we would conclude that Intel's stock is actually selling for much less than AMD's after we included the amount that Intel spends on buyback. However, there really is no correlation as stock can rise and fall based soley on speculation of future stock value rather than company performance.

Next the Reuters IW author, Scott Hillis, makes a comment:

That's because despite taking a beating in the market share battle, Intel is slashing costs, jettisoning underperforming businesses and launching new chips that are faster, more energy efficient and, crucially, more profitable.

The problem is that this statement is wrong. AMD's biggest liability was its Spansion flash memory business which it divested. Notice that there is no mention of this. In contrast, Intel still retains its own flash memory business. Again, no mention of this. The reality is that it is AMD that has slashed costs while Intel has trimmed. There is no mention of the fact that AMD's chips will be more energy efficient just one month later when it releases 65nm. There is no mention that AMD's K8L architecture will be considerably faster in 2007. Nor is there any mention of the way that AMD's costs will drop all during 2007 as production ramps up on the 300mm process with FAB 36. Finally, the notion that these chips are more profitable for Intel doesn't exactly fit with the previous statement about steep cuts in chip prices. The author is therefore either biased or extremely ignorant of the subject that he is trying talk about. This tends to be the problem where inconvenient facts tend to be rationalized away in favor of pro-Intel biases. There is no reason why anyone should believe arguments based soley on pro-AMD biases however most people are far more likely to encounter arguments based on Intel biases.

Saturday, October 14, 2006

What Quads May Come

I had a technical look at AMD's K8L in AMD Unveils Barcelona Quad-Core Details . There are improvements that I had not heard about before. There are several that should make a difference and boost both K8L's SSE and Integer performance. From this information it appears that 2007 will be much tougher for Intel and that it will lose most if not all of the ground it gained with the release of Core 2 Duo.

The K8L architecture seems to have more tweaks than I had first thought. There are many changes that pull AMD up to where Intel is now with Core 2 Duo. For example, Intel has had a dedicated stack unit since Pentium M and now AMD matches this with its own Sideband Stack Optimizer. Intel has been ahead of in out of order execution and AMD moves closer with enhanced reordering and some out of order loading. Intel had already increased the size of the reorder buffer.

AMD now has data dependent latency on the divide instruction as Intel had already done. AMD greatly improves SSE width and the number of SSE operations. It doubles the bandwidth of Prefetch and the L1 bus. Again, these are changes that Intel had made from Yonah (Core Duo) to C2D. AMD might have a small advantage here due to twin L1 buses rather than just a wider L1 bus.

It is also no secret that AMD's K8 architecture works better with lower latency memory like DDR. Intel has been pushing Jedec toward DDR2 and higher because its larger cache tends to reduce the bad effects of latency while it hurts AMD's smaller cache architecure more. However, AMD is now moving to a split memory controller where each half can act independently. This along with the larger L3 cache, improved memory scheduling, a better write burst mode, and out of order loads more than makes up for the higher latency of DDR2. We should see a big improvement for AMD with DDR2 and into DDR3. This again, moves AMD up to where Intel is now.

AMD's L3 is also more complex than I had thought. I had expected the L3 to be exclusive but AMD has apparently decided to make it more complex. They use it exclusively with data but can use it inclusively with instructions. This is a big improvement and would tend to negate most of Intel's advantage with inclusive cache and its advantage with the shared L2. However, this multi-mode function allows AMD to retain its current cache size flexibility. It would be able to change the size of both L2 and L3 without a severe penalty in performance. Intel does not have this flexibility. Further, Intel's use of shared L2 leaves the processor vulnerable to cache thrashing where one thread pushes instructions out of the L2 only to have its own instructions pushed out by the first thread in a cpu clock wasting tug of war.

AMD is expanding the number of FastPath instructions. This is important because the decoders can do 3 fastpath instructions or 1 complex instruction. The more instructions that are fastpath the better. This again makes up for some of the improvement that Intel gained by adding a fourth decoder and macro-ops fusion. This is also helped by doubling the Prefetch as Intel has done.

Overall, AMD moves much closer to Intel's C2D with the changes in K8L. AMD will add to its already greater memory access speed and match or beat C2D in terms of SSE. However, the changes don't quite add up to C2D's 4 instruction issue and Macro-ops fusion. I would guess that AMD will remove about 3/4 of the current IPC gap. This would mean that a K8L of 3.1 Ghz would match a C2D of 3.0 Ghz. This is much better than today where it would take a 3.5 Ghz K8 to match the same 3.0 Ghz C2D.

I am very curious to see how the testing will go when K8L is released as Antares for the desktop in Q3 07. There are so many improvements that move K8L closer to C2D that it will be difficult to bias the testing in Intel's favor. I'm guessing that when Intel is faced with the threat of K8L and the imminent loss of its performance advantages it will be sure to rush the new 45nm chips into reviewer's hands. Even these new chips though will probably not be enough to regain the C2D lead. I'm guessing that review sites like Tom's Hardware Guide and Anandtech will fall back on the classic cheat of comparing an overclocked Intel chip against a stock AMD chip. I can always hope that if the loss of Intel's lead becomes unavoidable that perhaps THG and Anandtech will even do some real testing with the second core (or other three cores on quads) loaded properly. They might even do some testing that actually stresses the memory bandwidth. I think that right now with the C2D lead that it is just too tempting for them to push the tests in Intel's favor and avoid real testing that would make C2D more even like loaded testing. It is generally expected that C2D will bog down under load more than K8. I'm still wondering how the review sites will avoid testing Barcelona (K8L Opteron quad) against Clovertown since Clovertown is expected to fair pretty badly in such a test. I guess we will see.

Monday, October 02, 2006

IDF, AMD, and Intel in Perspective -- 2nd Look

After a lot of discussion and consideration since the last article I've decided that FSB licensing does not mean that CSI has been scrapped. I had been under the assumption that CSI was essentially ready to go so the FSB licensing didn't make a lot of sense. I've now decided that I was overestimating Intel and that FSB licensing and Geneseo both make a lot of sense. I believe that my assumption that CSI was ready to go was incorrect and that, in fact, CSI will be two years late. CSI was supposed to be released in 2007 but now I doubt it will be released on any Xeon chip until 2009. In light of this, the FSB licensing would be a necessary stopgap. However, I do disagree with some assessments of Geneseo. Geneseo looks to me very much like CSI without cache coherency. If it were released today it would be somewhat competitive with HTX with the HT 3.0 standard. Geneseo/CSI would be competitive in terms of speed and latency however HT 3.0 includes a distance mode that can reach one meter, a power saving mode that is good for notebooks or reducing total power draw when full speed isn't being used, an unganged mode that splits a single 16 bit channel into two 8 bit channels, and HT 3.0 is hot pluggable. Secondly, by the time Geneseo is actually available we will probably be looking at HT 4.0 and Geneseo will be further behind. Nevertheless, it is a reasonable upgrade path for CSI to eventually catch up to HT. With a big upgrade to the Geneseo/CSI standard in 2009, CSI might be able to close the gap by 2010 or 2011.

The second issue is attitude. Intel and AMD have been complete opposites. Intel talked about photonics and the TeraFlop chip. Yet, Geneseo won't be ready for 2 - 2 1/2 years, TeraFlop is 5 years at the earliest, and photonics are indefinite. This is the height of vaporware. TeraFlop is unlikely to ever be released because it will most likely be outdated in five years. Similarly, it is not clear that photonics will ever be used for on-die signaling. The most likely use for fiberoptics would be something like a more robust HyperTransport link for connecting multiple cpu boards. This could even be used to connect a memory board to the processor board. The use of fiber would remove the need to run serpentine data traces all the way across the motherboard to the DIMMs. Memory could be on its own board and just plug in with fiber. I/O could also be on its own board. This would make boards easier to design because the signal path would be much smaller and easier to control. It would be easier to design just a cpu board or just a memory board or just an I/O board. I could see this application in servers. Interestingly, this would move servers back toward the way they used to be as minicomputers when the cpu occupied half a dozen to a dozen separate boards.

If Intel has been talking about technology in the distant (or nonexistent) future; AMD has been amazingly silent. AMD has not been merely quiet; it has been so secretive that even though both socket AM2 and F have been released there are still no pinout diagrams available. Typically, these pinout diagrams would be available a little before official release because chipset designers and motherboard makers would already have them. Yet, a search of the sea of information on the internet gets no hits for pinouts for either socket. And, if AM2 and F are a secret then 4X4 would make the CIA and the US military envious. I've seen demonstrations of future weapons and spy devices that won't be released for two or three years. 4X4 is due within 3 months but the technical detail available when written down would still leave space free on the back of a postage stamp. There are rumors that it will be socket F or socket AM2. There are rumors that it will support a cache coherent HTX slot or it won't. There are rumors that it will be used for workstations and support registered memory for increased reliability or that will only use unregistered DIMMs for higher speed. It still is not clear what will differentiate 4X4 and the new FX from the 2xxx series of Opterons. 4X4 could include 2 cache coherent HT links or maybe not.

One could take Intel's ramblings about future technology as a bluff because it doesn't want to admit the glaring gap in its server lineup for the next two and a half years. However, one could equally take AMD's deafening silence as an indication that it has nothing to compete with Conroe on the desktop for the indefinite future. However, I doubt everything is quite as bad as that. Intel does have a huge problem with Woodcrest. The current performance problem is small and will get fixed just as soon as Intel releases a registered DDR2 chipset because it is really FBDIMM that is making Woodcrest look bad. Dropping FBDIMM will improve latency with full DIMM sockets and reduce power draw. This should allow Woodcrest to pull noticeably ahead of Opteron. However, when it comes to 4-way, Woodcrest just isn't quite tall enough to reach the bar. This is a problem. However, IBM has the robust XA-64e (Hurricane) chipset for servers and this is probably enough to make a big tin server out of the currently wimpy Woodcrest. The only thing Woodcrest really needs are robust 4-way and 8-way connections. The need for this is even more desperate with Clovertown and its two dies crammed into one package. It should have this with XA-64e and I'm certain that IBM will be releasing heavy duty servers based on Woodcrest (Clovertown or Yorkfield), perhaps later in 2007. There is some possibility that this could even be intentional on Intel's part as Intel might wish to give Itanium some breathing room before X86 steals the markets above 4-way. Intel's statements suggest that it is all too aware that Itanium will get pushed higher and higher up the n-way ladder until it is no longer viable. The clock is ticking.

If IBM and DDR will be the saving grace for Woodcrest then K8L (Barcelona) would be the same for K8. By the time Intel gets a DDR chipset out the door for Woodcrest AMD will be releasing K8L. This processor should give K8 a huge push in FP/SSE performance and a smaller boost in Integer performance. If the rumors are true then AMD will be moving up the release of K8L Opteron to Q2 07 and the release of the desktop version to Q3 07. AMD knows that it will lose marketshare if it can't deliver something more competitive. While Woodcrest with DDR will leap ahead of Opteron it won't leap ahead of the dual core version of K8L Opteron, and Clovertown should find itself looking at tailights from quad core K8L. Opteron on HPC is where Intel wishes it were with Itanium or perhaps Woodcrest. AMD has three announced supercomputing wins (Sun, Cray, and IBM) for machines that will all be in the top 5 and has just upgraded the Oak Ridge machine. In contrast, there have been no new announcements for either Itanium or Woodcrest in the HPC arena. To add insult to injury, the purely Opteron Cray PetaFlop machine will probably rank 1st until it is displaced by the hybrid Cell/Opteron IBM machine. This would put Opterons in the number 1 and 2 systems with Sun's Opteron machine still in the top 5 (possibly even ranking 3rd). This isn't quite so bad for IBM because it has had the lead with Blue Gene L and will be putting its Cell machine into the number 1 spot and it's current Blue Gene machines will still be in the top 5. The only machine competitive with these is the proposed 3 PetaFlop Fujitsu machine which won't be ready until 2010 or 2011. It could potentially make it to 2nd on the list but there may be others by then.

IBM's XA-64e should keep the Woodcrest/Clovertown lineage from falling off the map on quad core/quad socket but it will be nearly impossible for it to match Opteron through 2008. The extra instuctions like PopCnt in 2007 will steal some of Itanium's thunder. In addition to the robust HT 3.0 added to K8L Opteron in 2007, AMD will bounce the technology even harder when it releases Direct Connect Architecture 2.0 in 2008 (Intel will still be waiting for Geneseo/CSI). This will give Opteron even greater connectivity with 4 HT links. This will be good for Sun and Dell (as well as HP and IBM if they pursue it) and will put pressure on the Intel based SGI, HP, and IBM offerings. However, IBM should sell enough Xeon systems to make a nice return on its XA-64e development. Likewise, the massive available memory on SGI's boxes should maintain a specialized market for their products. It would be nice to see how these products compare with proper load testing but I haven't seen this yet from any of the major review sites. It has yet to be done for even the dual core processors.

I think Intel's vaporware show and tell is a symptom of being ahead and trying very hard to make everyone believe that they have a solid strategy to stay ahead. The reality suggests that Intel has almost no hope of getting ahead in servers and is in serious danger of losing its advantage in notebooks. AMD now appears to be very serious about producing a proper low power chipset for notebooks with ATI. This could be a problem for Intel in 2008 when AMD moves to an integrated CPU/GPU for notebooks. Intel will need to work hard to keep from falling behind in power consumption which has been the trademark of the Centrino line. However, pressure will come much sooner than 2008 as AMD qualifies multiple chipsets and wireless adapters for OEM's. This is a sharp contrast to Intel's all or nothing Centrino position.

I think AMD's silence is related to the fact that while everything seems to be improving in terms of servers, notebooks, and corporate customers the desktop is still a problem. Intel gained massively by both increasing IPC and increasing clock. It would be impossible for AMD to increase the clock quickly enough with K8 to offset the IPC gains that Intel has. AMD needs K8L to increase IPC and needs 65nm to continue increasing clock. This gives AMD a fighting chance on the desktop in 2007 but doesn't help today. Today, AMD is clearly behind on the desktop and is very much aware of this. I believe that AMD has been quiet and extremely secretive because they realize that they have to make a strong showing when they announce something new. People who were expecting a jump in performance with Revision F were disappointed; AMD doesn't want another disappointment. AMD only has three things it can show. It can show that the power draw is less with 65nm. This will be helpful for notebooks and with large scale servers to be sure, but won't do much on the desktop.

AMD has 4X4 which should regain the FX position by essentially selling two chips for the price of one. However, for this to have any effect, AMD will need demos or benchmarks that can take advantage of four cores. Most games and benchmarks do not. AMD is going to have to carefully plan a presentation for 4X4. I suppose it is possible that they are talking to one or more software companies for some kind of endorsement. We can only wonder. However, whatever 4X4 does, it will be compared to Kentsfield and it is unlikely that the major review sites will do anything but superficial testing. I wouldn't be surprised with claims that Kentsfield is faster in everything than 4X4 when common sense would indicate that Kentsfield will bog down on any memory intensive task. This simply indicates the very sad and unreliable state of big review sites today as I've already talked about in my articles on Tom's Hardware Guide and Anandtech. The final thing that AMD has to show is the demo of K8L. This should give some indication of where AMD will be in Q2 07. However, K8L is unlikely to have the same IPC increase as Conrore which means that AMD will need a higher clock to match Conroe's performance. Considering that AMD is currently trailing in clock this will be a challenge even with 65nm. So, Intel's banter is a company pretending that being ahead on the desktop means that it is ahead everywhere and that it will stay ahead. AMD's silence is a company that knows it is behind on the desktop and needs maximum effect to get any notice. Intel has a fighting chance to stay ahead on the desktop in 2007 but should be facing increasing competition in notebooks, corporate accounts, and servers. AMD seems to be improving in notebooks, servers, and corporate accounts, but is going to have to work hard to catch up on the destkop in 2007.