EEMBC, an industry consortium that develops benchmarks for embedded software and hardware, today announced that the organization is seeking participants for a new Machine Learning working group. Group members will collaborate to develop EEMBC’s Machine Learning Benchmark Suite, which will identify the performance potential and power efficiency of processor cores used for accelerating machine learning jobs on clients such as virtual assistants, smartphones, and IoT devices.
“Until now, benchmarks have focused on training processes in the cloud, neglecting performance and power consumption measurements for cores running learning inference models on IoT edge devices, such as those used by Amazon Alexa, Apple’s Siri, and Google Cortana,” said Peter Torelli, EEMBC president and CTO. “Participants in our Machine Learning working group will not only help usher in this new and much-needed area of measurement, but also ensure meaningful and fair representation for their companies’ products.”
Chaired by Intel’s Ramesh Jaladi, the Machine Learning working group is currently defining the first proofs of concept. Participants include Analog Devices, ARM, AuZone, Flex, Green Hills Software, Intel, Nvidia, NXP Semiconductors, Samsung, STMicroelectronics, Synopsys, and Texas Instruments.
