If you think about it, that is roughly in line with Moore's law - twice as fast every 2 years so over 12 years 2 to the power of 6 = 64 times. We can't clock 64 times faster so parallel execution and clever architectures have to be used which use up the 64 times transistors to get the speed.I suspected the benchmarks you mentioned were related to Smalltalk. I see only a 58x improvement with some fairly well parallelized C code between the single-core Pi B and the quad-core Pi 5.
Of course there are faster processors available as RPi run behind the bleeding edge on the processor for cost reasons, but their memories are near state of the art then and now.
Statistics: Posted by MikeDB — Wed Aug 21, 2024 10:46 am