Controller
Model of the controller installed in the SSD.
The controller is a control circuit, which, in fact, ensures the exchange of information between the memory cells and the computer to which the drive is connected. The capabilities of a particular SSD module (in particular, read and write speed) largely depend on this particular scheme. Knowing the controller model, you can find detailed data on it and evaluate the capabilities of the drive. For simple everyday use, this information is usually not needed, but for professionals and enthusiasts (modders, overclockers) it can come in handy.
Nowadays, high-end controllers are produced mainly under such brands:
InnoGrit,
Maxio,
Phison,
Realtek,
Silicon Motion,
Samsung.
Cache memory
Buffer memory is a small chip on the SSD drive that acts as a data transit between the drive and the motherboard. In fact, it acts as a kind of intermediate link between the computer's RAM and the drive's own permanent memory. The buffer is used to store the most frequently requested data from the module, which reduces the access time to them — the information is sent from the cache, instead of being read from the magnetic media. Usually, the larger the buffer size, the higher the speed of the drive, all other things being equal. Also, drives with a large amount of buffer memory reduce the load on the processor.
Memory type
The type of the main memory of the drive determines the features of the distribution of information over hardware cells and the physical features of the cells themselves.
—
MLC. Multi Level Cell memory based on multi-level cells, each of which contains several signal levels. MLC memory cells store 2 bits of information. Has optimum indicators of reliability, power consumption and productivity. Until recently, the technology was popular in entry-level and mid-range SSD modules, now it is gradually being replaced by more advanced options in the manner of TLC or 3D MLC.
—
TLC. The evolution of MLC technology. One Flash Memory Triple Level Cell can store 3 bits of information. Such a recording density somewhat increases the likelihood of errors compared to MLC, in addition, TLC memory is considered less durable. A positive feature of the nature of this technology is its affordable cost, and various design tricks can be used to improve reliability in SSDs with TLC memory.
—
3D NAND. In a 3D NAND structure, several layers of memory cells are arranged vertically, and interconnections are organized between them. This provides greater storage capacity without increasing the physical size of the drive and improves memory performance due to shorter connections for each memory cell. In SSD drives, 3D NAND memory can use MLC, TLC or QLC chips - more details
...about them are described in the corresponding help paragraphs.
— 3D MLC NAND. MLC-memory has a multilayer structure — its cells are placed on the board not in one level, but in several "floors". As a result, manufacturers have achieved an increase in storage capacity without a noticeable increase in size. Also, 3D MLC NAND memory is characterized by higher reliability than the original MLC (see the relevant paragraph), at a lower manufacturing cost.
— 3D TLC NAND. "Three-dimensional" modification of the TLC technology (see the relevant paragraph) with the placement of memory cells on the board in several layers. This arrangement allows you to achieve higher capacity with smaller sizes of the drives themselves. In production, such memory is simpler and cheaper than a single-layer one.
— 3D QLC NAND. Quad Level Cell flash type with 4 bits of data in each cell. The technology is designed to make SSDs with large volumes widely available and finally retire traditional HDDs. In the 3D QLC NAND configuration, the memory is built according to a “multi-level” scheme with the placement of cells on the board in several layers. "Three-dimensional" structure reduces the cost of production of memory modules and allows you to increase the volume of drives without compromising their weight and size component.
— 3D XPoint. A fundamentally new type of memory, radically different from traditional NAND. In such drives, memory cells and selectors are located at the intersections of perpendicular rows of conductive tracks. The mechanism for recording information in cells is based on changing the resistance of the material without the use of transistors. 3D XPoint memory is simple and inexpensive to produce, and offers much better speed and durability. The prefix "3D" in the name of the technology says that the cells on the crystal are placed in several layers. The first generation of 3D XPoint received a two-layer structure and was made using a 20-nanometer process technology.Write speed
The highest speed in write mode characterizes the speed with which the module can receive information from a connected computer (or other external device). This speed is limited both by the connection interface (see "Connector"), and by the characteristics of the device of the SSD itself.
Write IOPS
The IOPS provided by the drive in write mode.
The term IOPS refers to the highest number of I / O operations that an SSD module can perform per second, in this case, when writing data. By this indicator, the speed of the drive is often evaluated; however, this is not always true. Firstly, the IOPS values of different manufacturers can be measured in different ways — by the maximum value, by average, by random write, by sequential write, etc. Secondly, the benefits of high IOPS become noticeable only with some specific operations — in in particular, the simultaneous copying of numerous files. In addition, in fact, the speed of the drive may be limited by the system to which it is connected. In light of all this, it is generally acceptable to compare different SSD modules by IOPS, but the real difference in performance is likely not to be as noticeable as the difference in numbers.
As for specific values, for the write mode with IOPS
up to 50K is considered relatively modest,
50 – 100K — medium,
more than 100K — high.
Read IOPS
The IOPS provided by the drive in read mode.
The term IOPS refers to the maximum number of I / O operations that an SSD module can perform per second, in this case, when reading data from it. By this indicator, the speed of the drive is often evaluated; however, this is not always true. Firstly, the IOPS values of different manufacturers can be measured in different ways — by the maximum value, by the average, etc. Secondly, the advantages of high IOPS become noticeable only with some specific operations — in particular, when copying numerous files at the same time. In addition, in fact, the speed of the drive may be limited by the system to which it is connected. In light of all this, it is generally acceptable to compare different SSD modules by IOPS, but the real difference in performance is likely not to be as noticeable as the difference in numbers.
For modern SSDs in read mode, an IOPS value of
less than 50K is considered a very limited indicator, in most models this parameter lies in the range of
50 – 100K, but there are also
higher numbers.
TBW
The abbreviation TBW stands for drive time between failures, expressed in terabytes. In other words, this is the total amount of information that is guaranteed to be written (rewritten) to this module. This metric measures the overall reliability and lifespan of a drive—the higher the TBW, the longer the device will last, all other things being equal.
Note that knowing the TBW and the warranty period, you can calculate the number of rewrites per day (DWPD, see the relevant paragraph), if the manufacturer did not specify these data. To do this, use the formula: DWPD = TBW / (V * T * 365), where V is the drive capacity in terabytes, T is the warranty period (years). As for specific numbers, there are a lot of drives on the market with a relatively low TBW —
up to 100 TB ; even these values are often sufficient for everyday use for a considerable amount of time. However, models with TBW at the level
of 100 – 500 TB are more common. Values of
500 – 1000 TB can be classified as "above average", and in the most reliable solutions this figure
is even higher.
M.2 cooling
The presence of a cooling radiator in the design of the M.2 form factor drive.
The heatsink is usually a metal plate attached to the drive board. It improves heat dissipation, which is especially important under high loads associated with handling large amounts of information. M.2 drives with
a heatsink are intended mainly for high-performance systems, in particular gaming ones.
There is also a special type of thin and light graphene radiators. They are glued to the surface of the M.2 SSD, covering the key areas (controller and memory chips) that generate the most heat. This allows heat to be distributed more evenly and its accumulation to be minimized.
We also note that M.2 radiators are found as equipment on motherboards. So if the drive itself does not have this function, you can choose
a “motherboard” with a radiator for it.
