Processors

Understand Intel's new server processor numbering scheme

Scott Lowe explains what you need to know about Intel's new numbering scheme for its Xeon-based processors.

Intel has once again decided to change the numbering scheme for its new Xeon-based processors. If you're familiar with Intel's Core i3, i5, i7 processor lines, you'll recognize that the company uses similar numbering for the Xeon server processors.

Figure A is an example of Xeon processor numbering. Below the image, I explain the various elements in the numbering scheme. Figure A

Brand: For the purposes of server processors, this will be Intel Xeon processor and simply identifies the overall brand - Xeon. Product line: Wayness, maximum CPUs in a node: How many sockets are supported in systems housing this processor? In this case, a single socket is supported. Socket type: The socket type field identifies the capabilities of the processor socket type, such as memory speeds and other items tied to the socket. The higher this number (I assume it ranges from 1 to 9), the more capabilities inherent in the socket type. Processor SKU: This is the processor part number. In my limited review of Intel's current part numbers and the processor SKU field, it looks like Intel is being mostly consistent in part number with regard to clock speeds -- at least within each product line. Here's a look at what I've pieced together for the current line of E7 offerings. The E3 line carries different values. Because there is not official guidance from Intel that this field will be consistent, take this information for what it's worth.
  • 03. 1733 MHz frequency/2400 MHz bus.
  • 07. 1867 MHz frequency/2400 MHz bus.
  • 20. 2000 MHz frequency/2900 MHz bus.
  • 30. 2133 MHz frequency/3200 MHz bus.
  • 37. 2667 MHz frequency/3200 MHz bus.
  • 50. 2000 MHz frequency/3200 MHz bus.
  • 60. 2267 MHz frequency/3200 MHz bus.
  • 67. 2133 MHz frequency/3200 MHz bus.
  • 70. 2400 MHz frequency/3200 MHz bus.
Power usage: In Figure A, you saw that the example has an L in the power usage area; this denotes that the part in question is a low power processor. When there is no L here, it's not a low power processor. Version: As new versions of a process are released, you'll see this field populated with v2, v3, and so on. For a processor's initial release, this field is blank; there is no v1 denotation.

My take

I really miss the good old days when the processor's numbering scheme included the speed as a part of the model number. However, when the megahertz and the gigahertz began to take on less meaning, this scheme started to make less sense, as some processors with lower clock speeds handily outperformed their seemingly faster brethren.

I wish Intel would consider including the core count somewhere in the new numbering scheme, especially as processor core counts continue to increase. Further, given that there are only a limited number of speeds at which processors actually operate, it seems like the product SKU could have been a bit more meaningful.

That said, as for Intel's new scheme, if they're consistent with all of the fields, there's hope yet. It's not perfect, but it does help, although I suspect that most admins will continue to actually research processor specs before making significant purchases.

About

Since 1994, Scott Lowe has been providing technology solutions to a variety of organizations. After spending 10 years in multiple CIO roles, Scott is now an independent consultant, blogger, author, owner of The 1610 Group, and a Senior IT Executive w...

2 comments
dogknees
dogknees

Wouldn't it be great if there was a clear indication of the number of cores and threads, the clock speed, and the maximum number of CPUs in a system. Still, one out of 4 isn't too bad!

Editor's Picks