Intel Xeon E-2176G Power Consumption
We wanted to post a few figures from our testing that show the real selling point of the chips, low power.
Idle is around 33W and maximum power consumption hits just under 114W in our test bed. The solid part about the Intel Xeon E-2176G is that it is able to power machines at under 1A in 120V racks. For the low-cost colocation world, this is a great option.
Note these results were taken using a 208V Schneider Electric / APC PDU at 17.7C and 72% RH. Our testing window shown here had a +/- 0.3C and +/- 2% RH variance.
Intel Xeon E-2176G Market Positioning
Thes chips are not released in a vacuum instead, they have competition on both the Intel and AMD sides. When you purchase a server and select a CPU, it is important to see the value of a platform versus its competitors.
Intel Xeon E-2176G v. Intel Alternatives
At under $362, the Intel Xeon E-2176G provides a lot of compute performance. Six cores are a major upgrade over 6-7 generations of Intel Xeon CPUs in this segment which all were four cores. You can read STH’s Looking back at Intel Xeon E3-1200 V1-V6 to the New Xeon E-2100 for the history from 2009/2011 to 2018 in the segment.
The Intel Xeon D-2100 series still has some major advantages. The Intel Xeon D-2100 series can utilize higher capacity RDIMMs, has more memory bandwidth, and has 10GbE NICs built-in. It is hard to discount that value. The same comparison would likely apply to the Intel Xeon Scalable family where one gets more clock speed with the Intel Xeon E-2176G but misses on the platform benefits.
Compared to the Intel Atom C3000 series, the Intel Xeon E-2176G does not have all of the same platform features but is much faster in single-threaded workloads. The larger and more robust compute cores (as seen in our GROMACS test for example), and higher frequencies make this an option for low-cost low power 1U servers.
The decision tree branch that will lead you to the Intel Xeon E-2176G over Atom and Intel Xeon D siblings is the Intel iGPU. The iGPU is not the focus of our CPU benchmark suite, but it includes transcoding hardware support that may be of interest to some deployments with supporting software. One can use BMC graphics for day-to-day management and offload transcoding to dedicated hardware logic not present in mainstream Intel Xeon CPUs outside of the Xeon E-2100 line. If your application can utilize video features or even transcoding offload for Quick Sync video, then the Intel Xeon E-2176G is the option to choose.
Intel Xeon E-2176G v. AMD EPYC
At the time of this writing, AMD EPYC does not have a real competitor to the Intel Xeon E-2176G for servers. One can use an AMD Ryzen but until we see platforms like the Tyan Tomcat EX S8015 hit the market, and from multiple vendors, AMD’s excellent compute performance with Ryzen 2 is lost by not having platforms with features like IPMI. If someone wanted to use a consumer platform without management, then Ryzen is an option, but the market for that is very small. A year and a half into the Ryzen adventure, the ecosystem is still not picking up the slack.
Final Words
There are a lot of benefits to the Intel Xeon E-2176G. The high clock speeds are excellent and the ability to utilize high all core turbo clocks is welcome. Limitations abound though. If you want over 64GB of RAM you will have to wait until the new 32GB ECC UDIMMs are out to get to 128GB. Even with future memory technology, it is less capacity than every other modern Intel server platform, including the Atom C3000 line. It is the only platform that does not have 10GbE capabilities either built into the chip or the chipset. Its ace is its iGPU for applications like transcoding as using dedicated offload is important for that niche.
In the Intel Xeon E-2100 range, the Intel Xeon E-2176G is an option at a price point. Configurations are going to be driven by precise budgetary constraints and how much a large server OEM like Dell EMC or HPE, or a large dedicated hosting provider can incrementally charge for this over an Intel Xeon E-2146G or E-2186G. What we can see for sure is that any hosting provider deploying these 6 core chips are going to have an immediate advantage over hosting companies still using the older Intel Xeon E3 generations. This is a massive upgrade in what had been a sluggish space.
No AVX-512 means you cannot benefit from optimizations from that codepath.
It’s unclear to me which OpenSSL version STH did use and if it included support for it; and which versions have been used in the past, to compare. From what I can tell using numbers of my servers with a D-15xx, STH’s neither supported it nor did STH use the more relevant to today’s market ‘ecdsap256’ for benchmarking signing and verifying .