True Speed Technology
True Speed : Sustained High Performance
True Speed provides enterprise, corporate, and OEM customers with a real-world performance advantage. Lite-On drives with True Speed have their performance indicators measured using realistic data sets and significantly offer highly stabilized long-term performance. This brief introduction will explain the importance of these to your company’s choice of SSD supplier and allow you to judge the relevance of True Speed to your specific SSD application.
With speed improvements being a major reason for the adoption of SSD, it is understandable that both manufacturers and purchasers focus on this aspect of the specification, mainly through direct measures such as sequential read, sequential write, and random write. For corporate purchasers especially, it is these numbers that allow judgment of ROI and value. However, these figures are often given based on tests using data made up just one repeated bit in other words the data being written to the drive is continuously just binary 1s or 0s. Clearly this test data is completely unlike the normal data that would be stored on the drive. In order to simulate realistic performance, the preferable method would be to issue specs based on random data testing, as is standard procedure at Lite-On. The importance of this can be seen from our test bed results. Using Crystal Disk Mark, we measured four competitors’ non-True Speed drives against two of Lite-On’s drives with True Speed. Our drives with True Speed are specifically designed to perform in real-world situations so the performance is comparable between linear and random data. Non-True Speed drives displayed an average of 16% reduction for sequential read, 58% for sequential write, and 44% for random write speeds. Before selecting a supplier or OEM partner Lite-On suggests that companies carry out their own testing with random data to obtain realistic performance information for comparison with True Speed.
The True Speed system is designed to provide highly stabilized long-term performance and to prevent the drastic write speed losses that can occur with SSDs. Our long-term testing has revealed that these losses can be up to 90% of the initial sequential write speed in non-True Speed SSD models currently on the market. To understand this phenomenon, a simple explanation is provided below.
Optimum write performance is always obtained when SSDs are new and unused — in a “clean” state. With a brand new SSD, all the writing process is simple, as blocks are ready to be written to and large continuous areas of blocks are available. As the SSD is used, the structure of the files and space on the disk becomes more complicated — files are changed and become spread out across more blocks, deleted files free up previously used space, etc. We can describe a disk in this state as being “dirty.” The change from clean to dirty state happens relatively quickly, particularly in enterprise situations, and can easily be seen in less than an hour of continuous testing. The dirty state requires the drive to perform additional operations to write data, slowing the writing process. If a drive is not designed to reduce problems associated with being in a dirty state, then the write process becomes very inefficient.
True Speed is designed to prevent dirty-state slowdown using a combination of elements, primarily advanced firmware. Working closely with Marvell, Lite-On’s in-house firmware team developed advanced garbage collection and our exclusive Instant Restore technology; together these constantly work to bring the drive’s performance back to near clean-state levels.
The effectiveness of True Speed technology can be measured in testing: Three widely available non-True Speed drives and two Lite-On manufactured drives with True Speed were benchmarked using Crystal Disk Mark. These clean drives were subjected to one hour of continuous random writing to simulate a relatively short period of use. In this dirty state, the drives were then tested again using the same Crystal Disk Mark tests. Results showed average performance loss of 60% for sequential write and 58% for random write for non-True Speed drives. The performance change for the drives with True Speed was near zero.
Therefore in SSD applications where performance stability is required or where high write speeds are desirable, we strongly recommend the use of a drive with True Speed.