AMD Zen 4 vs. Zen 4C Performance, Zen 4C Core Scaling With Ryzen 5 8500G

First up was looking at the build times while compiling the Linux kernel on the Ryzen 5 8500G. Here it's interesting to note that having one Zen 4 and one Zen 4C core (2 cores / 4 threads) yielded 85% the performance of having just two Zen 4 (non-C) cores enabled. Considering the Zen 4C cores with the 8500G clock up to 3.7GHz rather than 5.0GHz, this was a very good showing for the Zen 4C in code compilation workloads.

Having one Zen 4 and one Zen 4C core enabled led to a 30 Watt average (32 Watt peak) while compiling the Linux kernel compared to two Zen 4 cores leading to a 34 Watt average and 36 Watt peak. Enabling one Zen 4C core led to the CPU power consumption average going up by just three Watts. When looking at the CPU power consumption when scaling up the number of Zen 4C cores enabled, the power increase while compiling the Linux kernel was going up by about 3 Watts with each additional Zen 4C core enabled. In comparison, going from one to two Zen 4 (non-C) cores was a 7 Watt increase on average.

Going from one Zen 4 core configuration to also having one Zen 4C core online caused only a minor increase in thermals.

These performance and power/thermal findings were similar across various other codebases benchmarked for the code compilation speed across multiple jobs.