We test all SSDs in a manner that closely replicates a typical user scenario. This means we test the drive as our system disk filled 50% with data. We do it this way because we want to simulate what an actual user will be capable of getting in terms of performance. Testing empty secondary devices can never accomplish this because no one ever runs their SSD as an empty secondary device.
Running in our user state, we can hit advertised sequential read speeds without issue. However, as we've seen from day one of the E16 era, E16 SSDs can ONLY hit 4,400 MB/s sequential write speeds when they are empty or almost empty. In short, more evidence that supports why we test the way we do. That said, we are hitting nearly 4,300 MB/s seq write speeds. Additionally, Q1T1 random performance is top-notch.
Anvil's Storage Utilities
When evaluating SSDs, we always place more value on total read performance because it is the hardest to come by in the world of storage, and it matters most in relation to user experience. The FireCuda 520 easily outperforms Samsung's 970 Pro when reading data on an AMD Gen4 enabled platform. Intel's Optane 905P is the runaway victor, but that SSD will set you back $1,200.
Max IOPS is all but meaningless to the average user. However, we are showing what the drive can achieve in our user state anyway. 666K read and a whopping 869K write when 50% full.