DDR (Double Data Rate) memory is the standard that has replaced the traditional SDR SDRAM memories (better known as “SDRAM memories” or “DIMM memories”) and is very well received by the market, especially in the segment of personal computers . But what made it so well accepted? What are the characteristics that differentiate it from the previous pattern? What are its advantages? The answers to these and other related questions you check below.
Emergence of DDR memories
At a time when Intel’s Pentium III processor was one of the major products of its kind on the market, the standard FSB (Front Side Bus) rate – essentially the speed at which the processor communicates with RAM – was 133 MHz, equivalent to 1,064 MB per second. However, it is known that, as a rule, the motherboard chipset does not use the FSB frequency to communicate with the memory, but rather the speed of the memory. At that time, the standard for memory speed was also 133 MHz (known PC-133 SDRAM memories), which also provided a transfer rate of 1,064 MB per second. It is possible to note, therefore, that there was a certain “balance” in the communication speeds between the components of the computer.
However, with the release of chips such as Intel’s Pentium 4 and AMD’s Athlon, this “balance” ceased to exist as the FSB of the processors started to have more speed, while the memories remained in the PC- 133, keeping the frequency at 133 MHz. Under these conditions, this means that the computer as a whole can not take advantage of all processing capabilities.
For Pentium 4 users there was even an alternative: use Rambus (or RDRAM) memories. This type was faster than the PC-133 memories, but it had some drawbacks: it only worked with Intel processors, had a very high price and the motherboards that supported the Rambus memories were also very expensive.
At the same time, DDR memories were already a reality, but Intel tried to popularize Rambus memories, which made it “ignore” the existence of the first. AMD, in turn, needed an efficient alternative that could work fully with its new processors. The company ended up betting on the DDR memories and, from there, this type became popular, especially since Intel soon had to adhere to the idea.
But the simple emergence of DDR memories was not an immediate solution to the problems of speed between memories and FSB. Only with the release of the Dual-Channel DDR memories does the solution become effectively effective. The subject will be discussed later.
DDR memory operation
The DDR memories are quite similar to the SDR SDRAM memories. The latter work in a synchronized way with the processor, avoiding the problems of delay in previous technologies. The great differential of DDR technology, however, lies in its ability to perform twice as many operations per clock cycle (in a nutshell, the speed at which the processor requests operations). Thus, while a SDRAM PC-100 SDRAM works at 100 MHz, for example, a DDR module with the same frequency causes it to double, that is 200 MHz.
But how is this possible? In memories, the data is stored in spaces called cells. These are organized into a sort of matrix, that is, they are oriented in a scheme that resembles rows and columns. The crossing of a line with a column forms what we know as the memory address .
Typically, in read and write operations, you can only access one line at a time. But DDR memories have a “trick”: they access two different positions, but both on the same line. This is why this technology can perform the operations fold per cycle, one in the beginning and the other in the end.
Because of this feature, DDR memories now have a different naming pattern. SDRAM SDRAM modules include expressions such as PC-100 and PC-133, where the number indicates the frequency. Thus, a PC-133 comb informs that the device works at 133 MHz. In DDR memories, this also occurs, but considering the duplicity feature per cycle. So a DDR-200 module, for example, actually works at the rate of 100 MHz. But in the alternative nomenclature, such as PC-1600, for example, the amount of megabytes transferred per second is what is considered. Observe the table:
|PC-100 SDRAM||800 MB / s|
|PC-133 SDRAM||1,064 MB / s|
|DDR-200 or PC-1600||1,600 MB / s|
|DDR-266 or PC-2100||2,100 MB / s|
|DDR-333 or PC-2700||2,700 MB / s|
|DDR-400 or PC-3200||3,200 MB / s|
|Dual DDR-226||4,200 MB / s|
|Dual DDR-333||5,400 MB / s|
|Dual DDR-400||6,400 MB / s|
It is worth mentioning that these transfer values are theoretical, that is, they indicate the maximum range. In practice, a number of factors can influence the transfer speed. But, even being theoretical, how is this calculation done?
It’s simple: in their operations, the DDR memories can transfer up to 64 bits at a time, that is, 8 bytes. Simply multiply this value by the memory frequency plus the number of operations per cycle. Thus, the calculation of a DDR-400 module is as follows:
8 (64 bits) x 200 (frequency) x 2 (operations per cycle) = 3,200
The end result is given in megabytes per second.
Although very similar to the SDRAM SDRAM memories, the DDR memories have another considerable differential: they work with 2.5 V, against 3.3 V of the first. As such, they reduce power consumption, which is especially important for portable devices such as notebooks.
Physical aspects of DDR memories
Visually, it is easy to distinguish DDR memories from SDR SDRAM memories. The former have only one division in the module slot between the contact terminals, while the second has two. In addition, the DDR memories use 184 terminals, against 168 pins of standard SDR SDRAM.
Dual-Channel can be considered as a solution that softens the fact that the memories do not keep pace with the processors. To do this, the scheme causes the DDR memories to transfer twice as much data at a time. Thus, 3,200 MB per second can be make 6,400 MB per second.
This is possible because in the chipset of the motherboard – or even within processors, in the case of some more current models – there is a special circuit called memory controller , which accounts for all aspects of access and use of this. In Dual-Channel, this controller causes the DDR memories to transfer twice as much data at a time, that is, instead of 64 bits, they transfer 128 bits (16 bytes). With this, the calculation of the previous topic becomes:
16 (128 bytes) x frequency x 2 (operations per cycle)
To enable the Dual-Channel scheme on a computer, you must have a compatible chipset (or, if applicable, a processor). In addition, it is recommended to have one or two pairs of memory modules that are identical (or at least to the same specifications). Equality reduces the risk of problems. At this point, an interesting tip is to get a Dual-Channel kit, which offers two DDR memory combs to operate in this mode.
Refer to the motherboard manual to find out which slots the modules should be installed in order to activate Dual-Channel mode, as well as to know if any parameters need to be changed in the BIOS setup.
DDR memories have been widely accepted in the market, however, as the evolution of technology does not stop, especially for processors, new standards have had to be launched to keep up with the latest chip speeds: these are DDR2 and DDR3.