Random Read & Write Patterns
Now we’ll see the dependence between the drives’ performance in random read and write modes on the size of the data block.
We will discuss the results in two ways. For small-size data chunks we will draw graphs showing the dependence of the amount of operations per second on the data chunk size. For large chunks we will compare performance depending on data-transfer rate in megabytes per second. This approach helps us evaluate the disk subsystem’s performance in two typical scenarios: working with small data chunks is typical for databases. The amount of operations per second is more important than sheer speed then. Working with large data blocks is nearly the same as working with small files, and the traditional measurement of speed in megabytes per second becomes more relevant.
Let’s start with reading.
Intel’s SSDs are unrivalled at reading small random-address data blocks. OCZ’s Indilinx-based drives are only half as fast (and the Agility doesn’t differ from the others in this respect). The transition to the newer tech process has made Intel’s product somewhat better in this test than before. The OCZ Summit is the last one again: it is slower than the other OCZ drives on smallest data blocks. It looks like its controller cannot cope with processing such a huge number of very small requests.
The Summit joins the other SSDs from OCZ when processing large data blocks whereas the Agility falls out of that group. The Agility really seems to have fewer controller channels. As the data block size is getting larger, the OCZ drives catch up and even outperform the first-generation X25-M but Intel has left no chance to its opponents by releasing the updated model whose performance is unrivalled.
We’ve got a very interesting picture here. Intel’s drives are obviously better than their opponents but the first-generation model is faster at processing smallest data blocks. This must be due to changes in firmware. Hopefully, the difference is not a bug but the result of optimizations of the drive’s performance in other modes. Anyway, those 10 thousand operations per second is very, very good for an MLC-based solid state drive. The SSDs with other controllers cannot yet match this result. OCZ’s Indilinx-based products perform differently in this test as they have different firmware. The Vertex Mac Edition is the best of them, perhaps because its firmware version is higher. The OCZ Agility is slower in comparison with the others. It is a low-end product indeed. However, it is still faster than HDDs whereas the OCZ Summit is downright poor. If you compare its results with those of the Corsair P128, you can see that the Corsair is much better even though these SSDs share the same platform.
Both models from Intel hit against their maximum of 70-80MBps and give way to the Vertex Mac Edition which delivers a top write speed of 200MBps. Funnily, this is the only Indilinx-based drive that confidently copes with writing large random-address data blocks. The performance of the ordinary Vertex and the Turbo fluctuates wildly, obviously due to imperfect firmware. The Agility is stable and predictable but too slow.
The OCZ Summit is a disappointment again. Its writing performance is too low.