Intel's Lynnfield mysteries solved
At the Intel Developer Forum in San Francisco, the chipmaker cleared up the minor mysteries surrounding die photos of new Nehalem-based microprocessors.
The mysteries of the Lynnfield and Jasper Forest die photos (from last week's post titled "Investigating Intel's Lynnfield mysteries") were all cleared up at the Intel Developer Forum last week, and as expected, there was nothing sinister going on--just some confusion in Intel's graphics arts department.
With the help of the always-helpful George Alfs of Intel's press relations department and Intel vice president Mooly Eden (general manager of Intel's PC Client Group), we got everything straightened out. Literally!
Here's the die photo of Intel's Lynnfield chip from my previous post:
This is the newest (shipping) part based on the Nehalem microarchitecture, differing from the earlier Bloomfield by the addition of an on-die PCI Express controller. Both chips are made in Intel's 45nm process technology.
According to Eden, the Lynnfield chip design is shared with several other Intel chips that will be on the market soon, including the Core i7 Mobile processor code-named Clarksfield and a server processor code-named Jasper Forest.
Here's Clarksfield in a new photo that was used in many Intel presentations at IDF:
Like the Jasper Forest image in that earlier post, this Clarksfield photo is a mirror image of the Lynnfield shot. (The remaining difference in rotation is inconsequential, since the direction of "up" is just a matter of convention.) All of the images at IDF were consistent, so it's pretty clear the older Lynnfield picture was just reversed somehow before it was labeled.
It's easier to see in the Clarksfield photo that a portion of the chip, a narrow vertical strip at the top left corner, has no circuitry on it. Alfs explained to me that the QuickPath Interconnect (QPI) I/O circuitry in this area was simply deleted from the chip when it was taped out, since it wasn't going to be used in the Clarksfield or Lynnfield variants. This makes good sense, since if there's nothing there but blank silicon, a manufacturing defect in that area can't kill the chip.
Alfs went on to say that Intel chose not to create a new layout for the chip, even though it could have been smaller without QPI or the third memory channel (also unused but fully present on the chip), because the costs of the effort just wouldn't have been worth the savings. I'm a little surprised by this decision given that Clarksfield and Lynnfield are likely to sell many millions of units, but it's a reasonable explanation.
The Jasper Forest chip, on the other hand, has all of this circuitry present and active: three memory channels, a QPI port, and even some features not found (or at least not enabled) on the PC parts such as hardware acceleration for RAID 5 and RAID 6 storage and support for PCIe-based clustering.
Some interesting IDF presentations went into detail on how Jasper Forest can be used in embedded systems such as storage controllers and high-speed network routers.
As we learned at the Linley Tech Processor Conference earlier this month, Intel has seen strong demand for Nehalem-based chips in networking gear. With its higher level of integration than previous Nehalem server chips such as the Xeon 5500 series, Jasper Forest will be an even better fit for these systems.
My only remaining question about Intel's publicity images relates to the photo on Mooly Eden's biography page. I suspect it may be Photoshopped, since I've never seen Mooly looking that somber in real life...