Actually, expected rather than stellar would be a more appropriate adjective for the performance of apple's upcoming chip. Considering TSMC's N3 reported transistor density fluctuating between 40 and 60 percent higher than N5 and N4.
However, we need to see SOCs from competitors on the same node to make valid comparisons. Last time this happened in recent years was between snapdragon 8 gen 2 and a16 bionic.
On the same node the latest SOCs are for all intents and purposes about equal. Yes , a 16 bionic is about 25 percent faster on the cpu front.But at what cost? By being about 75 percent slower gpu wise and much slower on the neural engine front.
And of course what remains an open issue is sustained performance as mobile benchmarks have some small troubles capturing it. But this is an issue for another debate.
Point being that comparing SD 8 gen 3 with a17 bionic might be invalid based on the info we know until now.