Industry-Standard Benchmarks for Embedded Systems
EEMBC, an industry alliance, develops benchmarks to help system designers select the optimal processors and understand the performance and energy characteristics of their systems. EEMBC has benchmark suites targeting cloud and big data, mobile devices (for phones and tablets), networking, ultra-low power microcontrollers, the Internet of Things (IoT), digital media, automotive, and other application areas. EEMBC also has benchmarks for general-purpose performance analysis including CoreMark, MultiBench (multicore), and FPMark (floating-point).

IoTMARK


An EEMBC Benchmark 

IoT-Connect FAQs

Will EEMBC implement a 'very' sleepy node profile?

A very sleepy node may fit the profile of an energy harvesting device. We are not planning to support such nodes. There is nothing inherently wrong with the approach even for those types of nodes, but implementation wise it would be a different implementation. An energy harvesting device will not consume any energy off of the energy monitor, and as such will require a different way to measure the energy consumed.

With mesh networking on the horizon for Bluetooth, is EEMBC planning on including Bluetooth Mesh in its IoT Benchmark suite?

For the first release, the IoT Connect Benchmark working group is currently working on incorporating support for Bluetooth Smart (BLE). For subsequent releases, we are discussing Thread mesh networking. Support for these protocols will help define how mesh endnodes are benchmarked. Once Bluetooth Mesh is more mature, and public, our working group will discuss adding it to the IoT Connect benchmark as a new profile.

What types of wireless connectivity are supported by the IoT Connect Benchmark? Is wifi on the list?

The working group is currently defining benchmark profiles for Bluetooth Smart (not Bluetooth Mesh at this time). For the subsequent version, we are considering Wifi, Thread, and LoRA.

Why doesn't this IoT benchmark cover gateways and servers, since they are also part of the IoT?

Gateways and servers are covered by other EEMBC benchmarks. Scale-out servers are covered by EEMBC's ScaleMark (http://www.eembc.org/bigdata/about.php). Gateways are handled by the Gateway benchmark working group (http://www.eembc.org/iot-gateway/about.php)

For the IoTMark-BLE, does EEMBC specify an antenna, transmit power, or other transmission details?

No, the antenna and transmit power is left to the implementer to set. There is nothing to stop anybody using a top range antenna, and going as low as they can with the Tx power. But it does not matter if you use a good antenna and transmit at -30dBm or a lousy one and transmit at -20dBm, it will have negligible effect on the overall benchmark score. You can also put the receiver just next to the transmitter, hide it in a Faraday cage and transmit at -70dBm if your device is capable of it, it should still not make a difference to the score. Why? Once we get below 20dBm, it is unlikely we will be able to measure the power given everything else that is occurring in the system (0 dBm = 1mW Tx power, so -20 dBm is 1mW / 10^2 or 10uW). Remember - we are benchmarking the system under very specific operating conditions, and Tx power is not one of them. This includes CPU overhead in processing the BLE protocol stack, as well as other system-level functions.