Power Usage Benchmark with EEMBC

The latest benchmarks of wolfSSL power consumption on an STM32L476G device are up (https://www.eembc.org/viewer/?benchmark_seq=13436). What we found is that using wolfSSL’s SP math (with assembly speed ups) is superior on the device. It has a positive impact on both the speed and power consumption.

With the measurements used with EEMBC (https://www.eembc.org/) higher final scores are better. Without using any optimizations in building wolfSSL the power usage collected was 2170 and performance was 502. Once turning on optimizations and SP assembly the power usage was 13200 and performance was 3050.

The energy score is derived from an inverted, weighted, micro Joules per iteration. Similarly the performance is an inverted, weighted, microseconds per iteration. ECDSA operations saw a significant performance and power usage improvement with SP math enabled and assembly optimizations compiled in. ECDSA operations are the biggest resource consumers with TLS handshakes and a good indication of how long and how much power a TLS connection will use.

For information about wolfSSL and power usage contact facts@wolfssl.com.