It's an unannounced goal for the lowest power chips to get nearly identical performance in single thread.
Desktop chips typically use a ULT process vs XLT on the mobile ones. Ballpark 5% higher maximum frequency. But yes, single thread for mobile chips is converging/has converged with desktop ones. It's more a factor of power budget though. Currently cores target around 4-5W per core at TDP. That's quickly trending downwards towards 2-3W a core. No longer anything to gain by trying to shove 30W+ into it. This is also why mobile cores are competing with desktop ones. Converging per core power envelopes.
I don't see how 8 big cores and 8 little cores beats 16 big cores. Could happen, I guess. I wouldn't count on it.
While I think it's very unlikely in this case, there's a possible way. Throughput performance is heavily influenced by the efficiency of the core. A "stronger" core doesn't mean anything if it takes too much power to maintain its clocks. Lakefield is perhaps a good example. So
if Gracemont was exceptionally efficient even compared to Zen 3, it would have a chance.
Gracemont is good. Very good. But so is Zen 3. I think AMD will ultimately still have the upper hand in throughput vs Alder Lake, but perhaps by a smaller margin than some might think.