Micron 9100 MAX 2.4TB Enterprise PCIe NVMe SSD Review

128K Sequential Write/Read

We precondition the drive for 6,500 seconds, or 1.8 hours, receiving performance data every second. A sequential steady-state is achievable in a much shorter span of time than a random steady-state. We plot both MB/s and Latency. We plot MB/s using blue scatter and Latency using orange scatter. We observe that steady-state is achieved at 0 seconds of preconditioning, indicating that the previous 2x LBA fill phase achieved a sequential steady-state.

The 9100 MAX is performing well above its factory sequential steady state write specification of 2000 MB/s. We observe a small number of tightly bound outliers. However, even the outliers are all above factory sequential write specification. Average steady-state write performance at QD256 is approximately 2300 MB/s.

Micron's 9100 MAX shows off some more of its write prowess, again dominating the field from QD8-256. Overall write performance at this level is unprecedented as far as we know. Our combined write results to this point indicate that if even a small dose of write is mixed into a workload, the 9100 MAX will deliver best-in-class performance. The 9100 MAX easily outperforms its closest competitors; the DC P3608 and DC P3700. The XC100E5C shares a similar hardware configuration with the 9100 MAX however, the 9100 MAX outperforms XC100E5C by 750 MB/s.

Intel's dual controlled DC P3608 in RAID 0 mode easily wins this test. The 9100 MAX and the XC100E5C battle it out for second place at QD64-256. Overall though, the XC100E5C outperforms Micron's 9100 MAX with a pure 128K sequential read workload.

Conclusion (TL;DR): The similarly configured Techman XC100E5C slightly outperforms the 9100 MAX with pure sequential read workloads. The 9100 MAX takes the competition to the woodshed with a pure sequential write workload.