ATTO Disk Benchmark is perhaps one of the oldest benchmarks going and is definitely the main staple for manufacturer performance specifications. ATTO uses RAW or compressible data and, for our benchmarks, we use a set length of 256mb and test both the read and write performance of various transfer sizes ranging from 0.5 to 8192kb. Manufacturers prefer this method of testing as it deals with raw (compressible) data rather than random (includes incompressible data) which, although more realistic, results in lower performance results.
In these benchmarks we see some mixed results among the capacities. The 250GB and 500GB capacities didn’t reach their rated read speeds but passed their write ratings. On the other hand the 1TB had the opposite results of its smaller counterparts. When looking over the benchmark for the 250GB we see that there is a much bigger increase of the write speeds from the 4.0KB to the 256.0KB file size. In comparison, the 500GB scales relatively evenly through its test. However, notice that it only reaches its write rating at the very end. Finally, we have the 1TB capacity that, like the 500GB, only reaches a rating at the end of the test with the 8MB file size.
Crystal Disk Benchmark is used to measure read and write performance through sampling of random data which is, for the most part, incompressible. Performance is virtually identical, regardless of data sample so we have included only that using random data samples.
Here we have the 250GB benchmarks for compressed and uncompressed data. When comparing the compressed data on the left to the uncompressed on the right we see that there is an increase overall especially the write speeds. The SandForce controllers compress data on the fly and thus write speeds increase almost 170MB/s due to this feature. However, already compressed data will be transferred a bit slower, as can be seen on the left.
The 500GB model shows similar sequential speeds to the 250GB model, however we can see that there is a drastic drop in 4K write performance. This is due to the older architecture of the SF-2281 controller and its limited LBA mapping capability. In order to expand to 480GB or 500GB the controller needs to map 8KB LBAs rather than 4KB and thus there is a drop in 4KB performance. Besides this, everything else looks good.
We tested the 1 TB with only random/compressed data because the controller handles both types of data the same so there is no increase or decrease in performance. Overall, performance is what we expect to see out of a SM2246EN powered SSD. Let’s move on to our other tests!