We recently covered theĀ Intel Xeon E-2200 refresh launch. Just after that, we found that the launch was only for workstations. Instead, Intel will be re-launching the Intel Xeon E-2200 series for servers soon. That is, of course, only for some server applications. For servers that fall under the embedded category, the new Intel Xeon E-2200 embedded series CPUs are already out. Confused yet?
New Intel Xeon E-2200 Series Embedded SKUs Launched
Here is what we were able to get for information on the new Intel Xeon E-2200 embedded series CPUs.
We combined this with our value analysis but we had precious little information here to use.
There are a few features of the list we wanted to point out. The “M” seems to indicate that one can configure the TDP up or down. The Intel Xeon E-2254ME and Intel Xeon E-2276ME are listed as 45W TDP configurable down to 35W. Conversely, the low power “L” chip known as the Intel Xeon E-2276ML is listed as 25W but configurable up to 35W.
M and L are used in a somewhat strange way. Those same letter suffix notations in the context of other Intel Xeon CPUs, or theĀ Second Generation Intel Xeon Scalable CPUs mean increased RAM capacity.
Another strange suffix is the “G” which means integrated graphics on the Intel Xeon E-2100 and Xeon E-2200 series processors. Indeed, for the “G” processors here, they have integrated graphics as well. The Intel Xeon E-2276ML and Xeon E-2276ME SKUs drop the G, but retain Intel UHD Graphics P630 onboard.
Final Words
Luckily, the embedded market is a tight community. All of these naming convention inconsistencies are likely to be confined to a domain of a few individuals that know the parts intimately. Still, the new Intel Xeon E-2200 embedded series CPUs are another CPU launch ahead of the E-2200 server CPU re-launch we are eagerly awaiting.
Isn’t the top line missing in your first picture ?
We don’t know what these numbers mean.
Eric, you forgot the column headers in the chart.
And yes, Intel has pretty much been confused since the initial scalable launch. They need to get their marketting team in order and stop wasting money on multiple launches of the same products lol.
What is “Adj T*B/Price” in your Value-Analysis / how do you calculate it?
We are using a 30% uplift for a hyper-thread versus counting the SMT thread as a full 2x. That is generally around what we see in non-HPC workloads as a performance uplift from the additional threads.
Example E-2278GE:
C*B/Price: 3.3 * 8 / 694 = 0.0380 … check
T*B/Price: 3.3 * 8 * 2 (=2 threads per core) / 694 = 0.0760 … check
Adj T*B/Price: 3.3 * 8 * 1.3 (=+30% SMT bonus) / 694 = 0.0495 … not close to 0.0150 … i don’t get it.
And what about the E-2226GE … no SMT but C*B/Price does not match Adj T*B/Price?
Could you provide the math-expression you use for the ADJ T*B/Price calculation?
Fixed thanks.