Bulldozer, originally expected to ship in Q3 will now ship in Q3, but late Q3 instead of early. SemiAccurate is now hearing August for the desktop part, with a tickle of server variants in Q3, but no real volume until Q4. If the original roadmap was late June/early July, this is about a two month delay.
Why? There is a new spin. Performance of the parts is said to be ‘disappointing’ by several sources, but bear in mind that both sides have been sending out hamstrung samples of late to keep the rumour mill quiet. What they see is unlikely to be what we get. That said, the two explanations we hear are a bug fix and performance.
Performance is the obvious one, they found something that they can fix to go from disappointing to less disappointing in a step, but this seems unlikely. The other more detailed explanation is that there was a major bug that was expected to be fixed in the B1 -> B2 transition, but when silicon arrived, it wasn’t. If this is true, it will be a major issue for reasons other than the obvious. AMD was widely viewed to have put these type of systemic problems *COUGH*Barcelona*COUGH* behind them by now, and this will be a harsh wakeup call if true.
Some sources are describing the ‘shipping’ step as B3, others as C0. C0 would seem to fit the performance bump problem model, and B3 a bug fix. With the backdrop of sandbagging though, you can’t say for sure no matter what is printed on the chip.
Ivy Bridge on the other hand is far less clear. Sources all up and down the halls of Computex all say that the chip is delayed at least two months. The release date was recently moved from January 2012 to March 2012 with no explanation. CeBIT launch anyone? Given the die shots we have seen of Ivy, it looks like Intel is being far from honest about the benefits of their 22nm FinFET process, so this isn’t shocking.
Working silicon for Ivy wasn’t said to be back from the fabs until late November or early December, the delay isn’t all that shocking. It looks like Intel’s 1.3 year tick/tock cadence is running like a clooooock. In silicon, the more things change…….
Where does that leave things? AMD has some problems, but the delay of Bulldozer might free up 32nm SOI wafers for Llano. That is good. Llano is lower margin than Bulldozer, so that is bad. Ivy Bridge looks to be a fairly direct shrink of Sandy, so don’t expect a huge leap, that comes with Haswell, but not in the way you would expect. Between these two ‘issues’ and the 28nm slips, the rest of 2011 looks to be a pretty quiet year.S|A
Latest posts by Charlie Demerjian (see all)
- Thing go bump(gate) in the night for Nvidia’s GP100 Pascal GPU - May 3, 2016
- Cavium’s Octeon TX blends compute and packet moving - May 2, 2016
- Qualcomm releases Zeroth API to developers - May 2, 2016
- Another detail about Qualcomm server SoCs revealed - Apr 27, 2016
- AMD finally really honestly launches the dual Fury - Apr 26, 2016