I have generally considered the Fibonacci to be 'no math' because there's nothing more than add, subtract and compare operations. It's more a measure of stack handling and compiler optimisation but gives a ballpark insight for comparisons.This leads me to the observation that another benchmark is probably desirable that has no math.
I have found 'proper benchmarking' to generally be beyond my comprehension. Like most people I expect, I just want to know an averaged "Is it quicker ? By how much?". If something isn't substantially quicker it's a "nice, but nothing to write home about".
From your Fibonacci results it seems the Pico 2 ARM is about 25% faster than Pico which is good to hear, RISC-V about 30% faster.
I'm sure an ARM versus RISC-V comparison is going to be win some, lose some.
Not definitive but a good enough ballpark to work with for me.
Statistics: Posted by hippy — Sat Aug 17, 2024 6:09 pm