EEMBC Recognizes Leaders Driving Industry-Standard Benchmarks Supporting Key Technology Trends
Fifteen engineers honored for leadership in building benchmarks for IoT, Heterogeneous Compute, Ultra-Low Power, Scale-Out Servers, and Mobile Browsing
EL DORADO HILLS, Calif. — February 7, 2017— The 2016 results are in and the Embedded Microprocessor Benchmark Consortium (EEMBC) is honoring 15 individuals for their contributions to develop next-generation industry-standard benchmarks. Processor and microcontroller vendors are already using these benchmarks to deliver unbiased performance information to their customers - the system designers - and to design more effective controllers. System designers are using these benchmarks to evaluate different processing solutions as well as study the behavior of their own systems.
While many EEMBC members generously work for the industry’s benefit and the greater good, 15 engineers stand out for their contributions:
* Christophe Arnal (STMicroelectronics), Ben Boren (Intel), Craig Giglio (Silicon Labs), and Carlos Neri (NXP) were first to integrate and diligently test the beta version of the IoT-Connect[i] benchmark framework and provide valuable feedback that led to uncovering optimizations of the benchmark’s functionality and usability.
* Bob Martin (Microchip) led in developing the second generation of EEMBC EnergyMonitor, an integral component used to measure energy consumption in the EEMBC IoT-Connect, IoT-Security, and ULPBench-Peripheral Profile benchmarks.
* Mark Wallis (STMicroelectronics) co-chairs the IoT-Connect working group and is consistently leaping ahead to contribute very thorough and detailed engineering expertise on technical issues.
* Brent Wilson (Silicon Labs) co-chairs the IoT-Connect working group, keeps the extensive project focused, and provides comprehensive analysis of project challenges. Both Brent and Mark Wallis are also very active technical contributors to the EEMBC ULPBench program.
* Vlad Calina (NXP) led coding development of the EEMBC Heterogeneous Computing Benchmark’s[ii] prototype. His work saved the working group many man-months of effort.
* Rafal Malewski (NXP) chairs the Heterogeneous Computing Benchmark working group, shares his broad expertise on heterogeneous architectures, educates the industry, and manages the many details of this complex benchmark project.
* William Goh (Texas Instruments), Rafael Lajara (Analog Devices) and Monica Redon (Analog Devices) provided the analytic and empirical data on the next release of ULPBench[iii]-Peripheral Profile, and proving the validity of this benchmark suite.
* Stefan Schauer (Texas Instruments) has chaired the ULPBench working group since August 2013 and championed the release of ULPBench-Core Profile, and developed the specification and code for the ULPBench-Peripheral Profile.
* Shay Gal-On (Cavium) for chairs the ScaleMark[iv] working group, sharing his expertise in scale-out server applications and features, performing extensive testing, and contributing workload components.
* Cyril Bianconi (Intel) single-handedly created the BXBench 2.0[v] benchmark and has made it a more stable and comprehensive measurement tool.
“Creating meaningful performance benchmarks crosses company and product lines and challenges industry leaders to pool expertise, time, and efforts,” said Peter Torelli, EEMBC director of technology. “While the number of contributors to EEMBC continues to grow, these 15 key individuals have contributed to the EEMBC effort above and beyond their responsibilities within their respective companies, and with the support of these companies, have contributed to the greater good of the industry.”
# # #
EEMBC was formed in 1997 to develop performance benchmarks for the hardware and software used in embedded systems. EEMBC benchmarks predict the performance and energy consumption of embedded processors and systems in a range of applications (e.g. autonomous driving, mobile imaging, Internet of Things, scale-out servers, and mobile devices) and disciplines (processor core functionality, floating-point, multicore, and energy consumption).
EEMBC members include Ambiq Micro, AMD, Analog Devices, Andes Technology, ARM, C-Sky Microsystems, Cavium, Codeplay Software, Cypress Semiconductor, Dell, Flex Ltd., Green Hills Software, Huawei Technologies, IAR Systems, Imagination Technologies, Intel, Marvell Semiconductor, Microchip Technology, Nokia, Nordic Semiconductor, NVIDIA, NXP Semiconductors, Realtek Semiconductor, Redpine Signals, Renesas Electronics, Samsung Electronics, Silicon Labs, Somnium Technologies, Sony Interactive Entertainment, STMicroelectronics, Synopsys, Texas Instruments, and Wind River Systems.
[i] EEMBC IoT-Connect Benchmark is a suite of power efficiency benchmarks for IoT edge nodes, including both micro-controller and RF components, using multiple profiles to represent popular IoT applications and radio protocols. Battery life is of paramount importance in IoT applications.
[ii] EEMBC Heterogeneous Computing Benchmark is a system-level benchmark suite addressing performance requirements for applications such as autonomous driving and mobile imaging. This benchmark will support the optimal utilization of complex SoCs that include CPUs, GPUs, and hardware accelerators.
[iii] EEMBC ULPBench is a power-efficiency benchmark suite for ultra-low power microcontrollers. The working group has developed a ULPBench-Peripheral profile that includes interactions with the most commonly used peripherals including analog-to-digital conversion (ADC), serial port interface (SPI), real-time clock (RTC), and pulse-width modulation (PWM). Similar to the IoT-Connect benchmark, this benchmark suite is critical to identifying the low-power leaders in microcontrollers.
[iv] EEMBC ScaleMark benchmark measures the latency and throughput of scale-out servers running real workloads such as Memcached, Media Streaming, Origin Webserver, Load Balancing, and MySQL.
[v] EEMBC BrowsingBench provides a standardized, industry-accepted method to evaluate web-browsing performance and battery life on mobile devices and laptops.