ATTO is a test that operates outside of the file system installed on the host system. This allows the benchmark to test the performance of the drive without the limitations that some file systems can impose on the storage device. The typical file system does have some overhead and restrictions that can hamper readings of the sequential bandwidth in particular.
This approach gives a clear view of the sequential performance of the drive under testing and used by the majority of manufacturers to generate the specifications that they market.
The Proteus Plus easily meets its marketed specifications in the sequential testing with ATTO. The SSD performs well in the lower Queue Depths and then scales very nicely up to a maximum of 275MB/s in read speed and 263MB/s with the sequential write testing. These results are at the maximum that the SATA 2.6 interface can provide.
Crystal Disk Mark
Crystal Disk Mark measures a number of various types of file access patterns to the data storage device under testing. The benchmark uses both sequential and random data, along with varying Queue Depths for the random data, creating a compelling disk benchmark for users. There is also the ability to select between testing with compressible and incompressible data.
The Indilinx Barefoot controller performs the same regardless of the compressibility of the data at hand, so we will be using the standard default benchmark mode with Crystal Disk Mark.
The sequential read and write results are close to the results that we achieved with ATTO. The random 4K read at the Queue Depth of is 33.96, which is comparable with even current generation SSDs at that same Queue Depth. The write speed is considerably slower than today's models and the SSD does not scale well with the 4K write results at the higher QD. This is typical behavior expected from this generation of SSD.
The typical applications for the Proteus will typically not consist of heavy random write workloads.