Intel’s Microarchitecture Delay Analysis – Just How Much of the Schedule Shift is Real?
[Editorial] It would seem that what is probably the first time in many years, Intel’s Delays are following a never-before-seen pattern. It always troubled me that the ultra high end spectrum of the consumer processors (HEDT) was one microarchitectural generation (is that even a term?) behind. At any given time there were at least two architectures in the same consumer spectrum which is slightly odd. However after these delays, we might see three microarchitectures spanning the generations, which in my humble opinion is simply capitalistic insanity.
A Big Wibbly Wobly Mess of Concurrent Processor Architectures – Intel’s Roadmap Delay
So here is the problem, according to latest reports, Intel’s Skylake processors will be arriving in Q2 of next year (I don’t know whether its still valid anymore after the latest Broadwell-K delay or not) and Broadwell-K is actually slated for arrival after Skylake-S (which will consist entirely of locked processors) is released. So here we have our first oddity, an older architecture is slated for arrival after a newer generation? The obvious solution is ofcourse that the Broadwell-k delay to Q3 has also shifted the release of Skylake-S so that both the locked and unlocked siblings can release at the same time. Admittedly that would make slightly more sense. Basically around this same time, we would be looking at Haswell-E which is the successor to Ivy Bridge-E (Intel’s current HEDT spectrum) which as the name betrays is the Ivy Bridge architecture. So in Q2-Q3 2015 we will have Haswell-E on the 22nm node, Broadwell-E on the 14nm node and Skylake-S (which is a different architecture) on the 14nm node. Quoting something Dr.Who never said, its an absolute wibbly wobbly mess of Processor Microarchitectures.
Now to adding to the confusion is Intel’s market driven nomenclature of the HEDT processors which does not indicate that it its a generation behind. I also have a feeling Intel might mess with the nomenclature of Broadwell-K and Skylake-S some more to make it marketing friendly. So using current nomenclature and probable naming, an example would be the i7-5960X, the i7-6770 and the i7-5770K all existing at the same time in the market. And nomenclature be damned because if you follow the usual rules, you are going to be seriously misled. Ofcourse this won’t fool any of the tech savvy people out there but the non-tech savvy people will almost certainly opt for the ‘latest gen’ which, you guessed it, is the locked i7-6770. Enthusiasts like me would have to buy the older generation of Broadwell in order to opt for the unlocked multiplier, while the hardcore audience will be rewarded with an even older architecture. Hmm, I see a pattern emerging. Take a look at the picture below.
We just received a report from overclockers.ru who gave us very helpful insight. Apparently the Vietnamese plant is up and running and is already churning out test samples. It would appear that the Vietnamese plant will enter volume production very soon and the shift to Vietnam is almost certainly an effort to cut costs, which is a method of profit maximization. So with the PC Market more or less stagnant how do you make the most profit. See the normal people and OEMs are generally only interested in the latest generation, even if they have no idea whats going on inside. They will usually keep said product for quite some time without worrying about the new arrival. However enthusiasts would pay the premium to have an unlocked multiplier and a significant portion would do so for even one step of generation. Similarly, the hardcore enthusiasts would pay for the greater than 4 core count and they would pay for even one step of generation. An obvious product differentiation method begins to emerge to maximize profits. Juice the architecture from the average user first, then the enthusaist and so on.
Conclusion: So am I saying all the evidence adds up that this delay was artificial? Somewhat. Let me explain. Basically the Costa Rica plant going out of business and the transition to other plants would have caused an actual delay in the assembly line. However, you can bet your rigs that Intel is going to milk the delay for all its worth. The delay might have been natural to begin with, but I very much doubt that’s the case anymore. But there is only one thing that worries me. If this tactic is super successful, we might see Intel developing a very very unhealthy habit.
— Usman Pirzada (@usmanpirzada) July 4, 2014