How about the benchmarks isnt utilizing the SSD to the fullest just yet? Something like how the 8600GT destroys the 7900GTX in games that were released later.
Well as far as I know that benchmark doesn't exist yet so we'll have to go with what we have.
Before the benchmark changes the test procedure, I'd like to know if the new load is more or less representative for a mobile device workload. Average queue depth for PC client workloads is between 1-4, and I'm inclined to think mobile devices stay at even lower queue depth.
Also, even if we increase the queue depth, as I assume this was the main idea, UFS 2.0 also scales nicely with it. Image bellow is taken from
this presentation.
All that being said, the storage performance on the new iPhone is likely plenty fast for the usage pattern (lots of reads, few writes, most used apps are already in memory).