Looking at the outliers in terms of L1i MPKI makes me wonder how different the results might be under peak optimization with profile guidance, LTO, and BOLT. Does the "cactus" program have a hot region of code that's too large to effectively cache, or is it just poorly laid out by the toolchain?
Anandtech used to do SPEC benchmarks against all CPU from x86 Intel / AMD to Apple SoC. I don't know any site does this any more.
What's the practical impact of choosing eMAG as the baseline CPU? Does it alter the weighing of the components in the final score?