Now, having learned what it is and what and how it serves, many of you are probably thinking about purchasing a more powerful and productive RAM for your computer. After all, increasing the performance of a computer with the help of additional memory RAM is the simplest and cheapest (as opposed to, for example, a video card) method of upgrading your pet.
And ... Here you are standing at a display case with packages of RAMs. There are many of them and they are all different. Questions arise: What kind of RAM should you choose?How to choose the right RAM and not miscalculate?What if I buy a RAM, and then it will not work? These are reasonable questions. In this article I will try to answer all these questions. As you already understood, this article will take its rightful place in a series of articles in which I wrote about how to choose the right individual computer components, i.e. iron. If you haven't forgotten, it included articles:
—
—
—
This cycle will be continued further, and at the end you will be able to assemble for yourself a super computer, perfect in every sense 🙂 (if, of course, finances allow :))
Until then learning to choose the right memory for your computer.
Go!
Random access memory and its main characteristics.
When choosing RAM for your computer, you must definitely build on your motherboard and processor, because RAM modules are installed on the motherboard and it also supports certain types of RAM. Thus, the relationship between the motherboard, processor and RAM is obtained.
Learn about what kind of RAM does your motherboard and processor support? can be found on the manufacturer's website, where you need to find the model of your motherboard, and also find out what processors and RAM for them it supports. If you don't, it turns out that you bought a super modern RAM, but it is not compatible with your motherboard and will gather dust somewhere in your closet. Now let's go directly to the main technical characteristics of RAM, which will serve as a kind of criteria when choosing RAM. These include:
Here I have listed the main characteristics of RAM, which are worth paying attention to first of all when buying it. Now we will open each of them in turn.
RAM type.
Today the most preferred type of memory in the world is memory modules DDR(double data rate). They differ in release time and, of course, in technical parameters.
- DDR or DDR SDRAM(translated from English. Double Data Rate Synchronous Dynamic Random Access Memory - synchronous dynamic memory with random access and double data rate). Modules of this type have 184 contacts on the bar, are powered by a voltage of 2.5 V and have a clock frequency of up to 400 megahertz. This type of RAM is already outdated and is used only in old motherboards.
- DDR2 Is a widespread type of memory at this time. Has 240 pins on the printed circuit board (120 on each side). Consumption, unlike DDR1, is reduced to 1.8 V. The clock frequency ranges from 400 MHz to 800 MHz.
- DDR3- the leader in productivity at the time of this writing. It is not less widespread than DDR2 and consumes 30-40% less voltage than its predecessor (1.5 V). Has a clock frequency of up to 1800 MHz.
- DDR4- a new, super modern type of RAM, which is ahead of its counterparts both in performance (clock frequency) and voltage consumption (which means it has less heat). Support for frequencies from 2133 to 4266 MHz is announced. At the moment, these modules have not yet entered mass production (they promise to release them into mass production in mid-2012). Officially, fourth generation modules operating in DDR4-2133 at a voltage of 1.2 V were presented at CES, by Samsung on January 04, 2011.
The amount of RAM.
I will not write much about the amount of memory. Let me just say that it is in this case that size matters 🙂
All a few years ago, 256-512 MB of RAM satisfied all the needs of even cool gaming computers. At the present time, for the normal functioning separately, only the windows 7 operating system requires 1 GB of memory, not to mention applications and games. There will never be extra RAM, but I'll tell you a secret that 32-bit windows uses only 3.25 GB of RAM, even if you install all 8 GB of RAM. You can read more about this.
The dimensions of the strips or the so-called Form factor.
Form - factor- these are the standard sizes of the RAM modules, the type of construction of the RAM strips themselves.
DIMM(Dual InLine Memory Module is a double-sided type of modules with contacts on both sides) - mainly designed for desktop stationary computers, and SO-DIMM used in laptops.
Clock frequency.
This is a pretty important technical parameter of RAM. But the motherboard also has a clock frequency and it is important to know the operating frequency of the bus of this board, since if you bought, for example, a RAM module DDR3-1800, and the slot (connector) of the motherboard supports the maximum clock frequency DDR3-1600, then the RAM module as a result will operate at a clock frequency of 1600 MHz... In this case, all sorts of failures, errors in the operation of the system, etc., are possible.
Note: Memory bus frequency and processor frequency are completely different concepts.
From the above tables, it can be understood that the bus frequency multiplied by 2 gives the effective memory frequency (indicated in the "chip" column), i.e. gives us the baud rate. The name also tells us about it. DDR(Double Data Rate) - which means double data transfer rate.
For clarity, I will give an example of decoding in the name of the RAM module - Kingston / PC2-9600 / DDR3 (DIMM) / 2Gb / 1200MHz, where:
- Kingston- manufacturer;
- PC2-9600- the name of the module and its bandwidth;
- DDR3 (DIMM)- type of memory (form factor in which the module is made);
- 2Gb- the volume of the module;
- 1200MHz- effective frequency, 1200 MHz.
Bandwidth.
Bandwidth Is a memory characteristic that affects system performance. It is expressed as the product of the system bus frequency by the amount of data transmitted per clock cycle. Bandwidth (Peak Data Rate) is a complex measure of capability RAM, it takes into account data transmission frequency, bus width and the number of memory channels. Frequency indicates the potential of the memory bus per clock cycle - higher frequencies can transfer more data.
The peak rate is calculated using the formula: B = f * c, where:
B - bandwidth, f - transmission frequency, c - bus width. If you use two channels for data transmission, we multiply everything received by 2. To get a figure in bytes / s, you need to divide the result by 8 (since there are 8 bits in 1 byte).
For better performance RAM bus bandwidth and processor bus bandwidth must match. For example, for an Intel core 2 duo E6850 processor with a 1333 MHz system bus and a bandwidth of 10600 Mb / s, you can install two modules with a bandwidth of 5300 Mb / s each (PC2-5300), in total they will have the system bus bandwidth (FSB) equal to 10600 Mb / s.
Bus frequency and bandwidth are designated as follows: " DDR2-XXXX" and " PC2-YYYY". Here "XXXX" stands for the effective memory frequency and "YYYY" for the peak bandwidth.
Timings (latency).
Timings (or latency)- these are the signal time delays, which, in the technical characteristics of the RAM, are written in the form " 2-2-2 " or " 3-3-3 " etc. Each digit here represents a parameter. In order, it is always “ CAS Latency"(Working cycle time)," RAS to CAS Delay"(Full access time) and" RAS Precharge Time"(Pre-charge time).
Note
So that you can better understand the concept of timings, imagine a book, it will be our RAM, which we refer to. Information (data) in a book (RAM) is divided into chapters, and chapters consist of pages, which in turn contain tables with cells (such as in Excel tables). Each cell with data on the page has its own coordinates vertically (columns) and horizontally (rows). The RAS (Raw Address Strobe) signal is used to select a row, and the CAS (Column Address Strobe) signal is used to read a word (data) from the selected row (i.e., to select a column). A full reading cycle begins with the opening of the "page" and ends with its closing and recharging, because otherwise the cells will be discharged and the data will disappear. This is how the algorithm for reading data from memory looks like:
- the selected "page" is activated by the RAS signal;
- data from the selected line on the page is transmitted to the amplifier, and a delay is required for data transmission (it is called RAS-to-CAS);
- a CAS signal is given to select (column) a word from that row;
- data is transferred to the bus (from where it goes to the memory controller), while there is also a delay (CAS Latency);
- the next word goes without delay, since it is contained in the prepared line;
- after the call to the row is completed, the page is closed, the data is returned to the cells and the page is recharged (the delay is called RAS Precharge).
Each digit in the designation indicates how many bus clock cycles the signal will be delayed. Timings are measured in nano seconds. The numbers can range from 2 to 9. But sometimes a fourth is added to these three parameters (for example: 2-3-3-8), called “ DRAM Cycle Time Tras / Trc”(Characterizes the speed of the entire memory chip as a whole).
It happens that sometimes a cunning manufacturer indicates only one value in the characteristics of the RAM, for example, “ CL2"(CAS Latency), the first timing is equal to two clock cycles. But the first parameter does not have to be equal to all timings, and maybe less than others, so keep this in mind and do not fall for the manufacturer's marketing ploy.
An example for clarity of the effect of timings on performance: a system with memory at a frequency of 100 MHz with timings 2-2-2 has approximately the same performance as the same system at a frequency of 112 MHz, but with delays of 3-3-3. In other words, depending on latency, the difference in performance can be up to 10%.
So, when choosing, it is better to buy memory with the lowest timings, and if you want to add a module to the already installed one, then the timings of the purchased memory must match the timings of the installed memory.
Memory operation modes.
The RAM can work in several modes, if, of course, such modes are supported by the motherboard. it single-channel, two-channel, three-channel and even four-channel modes. Therefore, when choosing RAM, you should pay attention to this parameter of modules.
Theoretically, the speed of the memory subsystem in the dual-channel mode increases 2 times, in the three-channel mode - 3 times, respectively, etc., but in practice, in the dual-channel mode, the performance gain, in contrast to the single-channel mode, is 10-70%.
Let's take a closer look at the types of modes:
- Single chanell mode(single-channel or asymmetric) - this mode is enabled when only one memory module is installed in the system or all modules differ from each other in terms of memory size, operating frequency or manufacturer. It doesn't matter in which slots and memory to install. All memory will run at the speed of the slowest memory installed.
- Dual Mode(two-channel or symmetric) - the same amount of RAM is installed in each channel (and theoretically, the maximum data transfer rate doubles). In dual-channel mode, memory modules work in pairs: 1st with 3rd and 2nd with 4th.
- Triple Mode(three-channel) - the same amount of RAM is installed in each of the three channels. Modules are selected in terms of speed and volume. To enable this mode, modules must be installed in slots 1, 3 and 5 / or 2, 4 and 6 slots. In practice, by the way, this mode is not always more productive than the two-channel mode, and sometimes even loses to it in the data transfer rate.
- Flex Mode(flexible) - allows you to increase the performance of RAM when installing two modules of different sizes, but the same operating frequency. As in the dual-channel mode, memory cards are installed in slots of the same name on different channels.
Usually the most common option is dual channel memory.
To work in multichannel modes, there are special sets of memory modules - the so-called Kit memory(Kit-set) - this set includes two (three) modules, of the same manufacturer, with the same frequency, timings and type of memory.
Appearance of KIT-kits:
for two-channel mode
for three-channel mode
But the most important thing is that such modules are carefully selected and tested by the manufacturer itself, for operation in pairs (triplets) in two- (three-) channel modes and do not imply any surprises in operation and configuration.
Module manufacturer.
Now on the market RAM manufacturers such as: Hynix, amsung, Corsair, Kingmax, Transcend, Kingston, OCZ…
Each company has its own product for each product. marking number, according to which, if you decipher it correctly, you can find out a lot of useful information about the product. For example, let's try to decipher the labeling of the module. Kingston families ValueRAM(see image):
Decoding:
- KVR- Kingston ValueRAM i.e. manufacturer
- 1066/1333 - working / effective frequency (Mhz)
- D3- memory type (DDR3)
- D (Dual) - rank / rank... A dual-rank module is two logical modules soldered on one physical channel and alternately using the same physical channel (needed to achieve the maximum amount of RAM with a limited number of slots)
- 4 - 4 DRAM memory chips
- R - Registered, indicates stable operation without failures or errors for as long as possible a continuous period of time
- 7 - signal delay (CAS = 7)
- S- thermal sensor on the module
- K2- a set (kit) of two modules
- 4G- the total volume of the whale (both planks) is 4 GB.
I will give another example of marking CM2X1024-6400C5:
The marking shows that it is DDR2 module volume 1024 MB standard PC2-6400 and delays CL = 5.
Stamps OCZ, Kingston and Corsair recommended for overclocking, i.e. have overclocking potential. They will have low timings and a clock rate margin, plus they are equipped with heatsinks, and some even coolers for heat dissipation. during acceleration, the amount of heat increases significantly. The price for them will naturally be much higher.
I advise you not to forget about fakes (there are a lot of them on the shelves) and buy RAM modules only in serious stores that will give you a guarantee.
Finally:
That's all. With the help of this article, I think you can no longer make a mistake when choosing RAM for your computer. Now you can choose the right RAM for the system and increase its performance without any problems. Well, for those who buy RAM (or have already bought it), I will devote the following article, in which I will describe in detail how to properly install RAM into the system. Do not miss…
Best RAM 2019
Corsair Dominator Platinum
The best memory among classmates with high performance and innovation in RGB technology. DDR4 standard, 3200MHz speed, default timings 16.18.18.36, two 16 GB modules. The strips feature bright Capellix RGB LEDs, an advanced iCUE program, and Dominator DHX heatsinks. The only problem is that the height of the module may not be suitable.
Corsair, as always, surpasses itself with each new model, the Dominator Platinum is no exception. Today it is the favorite DDR4 memory set for gamers and powerful workstation owners. The appearance of the modules is sleek and stylish, appealing to gaming enthusiasts, DHX cooling works efficiently, and the performance of the planks is already ready to become a legend. In any case, it will provide the user with flagship parameters for many years to come. The memory now has a new design, a new, brighter 12-LED Corsair Capellix backlight. The (proprietary) iCUE software provides flexible memory tuning for maximum performance. If you have changed the motherboard or processor, and maybe a graphics accelerator, memory can be configured as native for any new component.
The memory's price tag is slightly higher than that of other manufacturers, but this is offset by the highest quality and amazing performance.
The article is constantly updated. Last updated 04/01/2013 p.
Random access memory (RAM)- this is a special memory (random access memory), which temporarily stores data and commands necessary for the processor to perform operations, and the access time to this memory (for the processor) does not exceed one cycle.
Data transfer to / from RAM is performed directly through the processor's ultra-fast cache memory (L2 or L3).
Timing (latency) of RAM is the time delay of the data exchange signal, i.e. this is the short time delay for memory "responsiveness" to data I / O. The memory performance directly depends on the timings and, as a result, the performance of the entire system is very dependent.
Timings are indicated on the memory modules in the form: 4-4-4-12, 6-6-6-18, 9-9-9-27 or as part of the CL4, CL5, CL9 memory module marking.
The first step in choosing RAM is your choice of motherboard and processor.
Since the memory is directly installed in the motherboard and the type of memory will depend on the mat. boards.
We wrote about this:
And the processor will work directly with the installed RAM, and the new processors have a built-in controller for exchanging data with the RAM.
About it here:
Memory type.
The following types of memory are used in desktop computer systems:
DDR(double data rate) - currently this type of memory is outdated and almost never used. The module has 184 contacts. Standard supply voltage 2.5 V.
Marked as PC-2700 333 Mhz, PC-3200 400 Mhz.
Since this type of memory has long been discontinued, we will not focus on it.
DDR2- This is a widespread type of memory at the moment. DDR2, unlike DDR, allows you to sample 4 data bits per clock at once (4n-prefetch), DDR only 2 bits per clock (2n-prefetch), i.e. DDR2 is capable of transferring 4 bits of information from the cells of the memory microcircuit to the input-output buffers in one cycle of the memory bus. The module is designed as a printed circuit board with 240 contacts (120 on each side) and has a standard supply voltage of 1.8 V.
Marked as PC-5300 667 Mhz, PC-6400 800 Mhz, PC-8500 1066 Mhz.
This type of memory is now widely used in desktop office and gaming computers. Due to the high frequency, low timings (latencies) and double the sampling rate, the memory shows high performance results.
DDR3- A new and no less common type of memory. DDR3 - allows you to sample 8 bits of data per clock (8n-prefetch). The module, like DDR2, is made in the form of a 240-pin board (only the key / slot is displaced and you cannot install DDR3 in the DDR2 slot), and the standard supply voltage is only 1.5 V.
Marked as PC-10600 1333 Mhz, PC-12800 1600 Mhz, PC-14400 1800 Mhz, PC-15000 1866 Mhz, PC-16000 2000 Mhz.
At the moment, this type of memory displaces DDR2 from new systems and will completely replace it in the future. DDR3 has found application only in gaming and overclocking systems, but is also fully implemented in multimedia systems and laptops. Since it has higher operating frequencies and much higher bandwidth in comparison with DDR2.
The power consumption of DDR3 memory is approximately 40% less than that of DDR2 memory, which is very important for notebooks and mobile systems.
For new systems, purchasing DDR2 memory is no longer cost-effective. Is that for office computers based on an earlier one with an integrated graphics core.
And when purchasing new components for home and gaming-overclocking systems, at the moment, you need to focus on DDR3. Since, all new mat. boards and new processors only support DDR3.
The only thing to consider is that DDR3 has slightly higher timings compared to DDR2, but due to its higher frequency and lower power consumption, it is the best choice for desktop and mobile systems.
Memory frequency.
The arithmetic here is simple: the higher the frequency, the more efficient the memory.
The main thing is that your motherboard supports the memory frequency you have chosen.
But do not forget that with increasing frequency, timings (delays) also increase.
The golden mean in DDR3 is 1600 Mhz with CL7 or CL8 timings.
For DDR2, the optimal frequency is 1066 Mhz with CL5 timings.
Timings.
Timings (delays) are in other words memory latency. That is, the speed of memory "responsiveness" is determined by the timings.
It turns out that the lower the timings, the faster the memory.
DDR had standard CL3 timings (3-3-3-9) at 400 MHz
DDR2 standard CL6 timings (6-6-6-18) at 800 MHz
DDR3 has CL9 (9-9-9-27) timings at 1600 Mhz
But there are modules with reduced timings / increased performance. Such modules are a little more expensive than standard ones, but they can significantly speed up the system.
They are sometimes called overclocking memory.
You can buy memory, the timings of which, for example, at the same frequency for DDR2 800 Mhz are only CL4 (4-4-4-12), and for DDR3 1600 Mhz - CL7 (7-7-7-21).
The only thing is that to ensure such a mode of operation, some manufacturers indicate the supply voltage of their chips higher than the nominal.
Dual-Triple-Channel memory and KIT memory.
The two-channel mode began to be used relatively recently. And the three-channel one is based only on the X58 gaming chipset of the LGA 1366 platform for the Core i7.
Dual-channel mode is a mode of operation of RAM, in which memory modules work in pairs, that is, the 1st with the 3rd, and the 2nd with the 4th (in the three-channel - "triplets" 1-3-5, 2-4 -6), and each pair is on its own channel - while in single-channel mode all memory modules are served simultaneously by one controller (so to speak, they work in one channel).
The total amount of available memory in a three- or two-channel mode (as well as in a single-channel mode) is equal to the sum of all the volumes of the installed memory modules.
The two-channel robots memory mode gives a very good performance boost. In theory, this mode doubles the memory bandwidth. In practice, the increase in dual-channel versus single-channel is from 10% to 70% (depending on the application).
Well, for the three-channel, the gain is insignificant so far, in comparison with the two-channel, only a couple of percent.
The memory of one volume, one frequency, one manufacturer, one type will work in three-two-channel mode. And it is also necessary that the motherboard and processor support this mode of operation. You can read about this in the articles:
But sometimes there are exceptions.
Two (three) completely identical memory modules (frequency, timings, size, manufacturer, type, and even from the same batch) may "refuse" to work in Dual Channel (Triple Channel) and lead the system to a blue screen.
It's like a lottery who is lucky enough to launch two or three regular modules, and who is not.
And you can't make any claims under the guarantee, since they work perfectly individually and in single-channel mode.
For ease of launching the memory into the DualChannel mode, motherboard manufacturers "paint" the memory slots of one channel with one color, the other with another. Accordingly, to make the memory work in dual-channel mode, you need to install modules in slots of the same color (more precisely, we read the instructions of the motherboard).
(an exception is when there are only two slots on the board, then you can check the number of channels with the CPU-Z program)
Single-channel memory operation mode is when the memory is inserted into adjacent slots (of different colors):
Dual-channel mode, memory is installed in pairs 1-3, 2-4 (in slots of the same color):
IMPORTANT!!! If you, on the motherboard, support dual-channel memory robots, and the memory is inserted into the 1st and 3rd slots (for example, 2 pcs. 1Gb each) and you decide to deliver the third bar to the 2nd or 4th slot ( let's say the same bar with a volume of 1Gb). Then you will "lose" the dual-channel memory mode, and the controller will go into single-channel.
The increase from the added memory will not be high, and from the loss of the two-channel mode, the performance will decrease somewhat.
To save the two-channel mode, add memory in pairs !!!
Three-channel mode, memory is installed in "triplets" 1-3-5, 2-4-6 (also in slots of the same color):
There are special sets of memory modules for work in multichannel modes.
So-called Kit memory(Kit-set) - this set includes two (three) modules, of the same manufacturer, with the same frequency, timings and type of memory.
But the most important thing is that such modules are carefully selected and tested by the manufacturer itself, for operation in pairs (triplets) in two- (three-) channel modes and do not imply any surprises in operation and configuration.
Appearance of KIT-kits:
for two-channel mode
for three-channel
In addition, these memory modules are equipped with passive cooling heatsinks, the presence of which allows the chips to cool themselves.
This is an indisputable plus and has a positive effect on the stability of the memory.
Based on the performance gain tests, the optimal choice for all systems (including office ones) is the Dual Channel memory mode.
Dual channel performance tests:.
That is, for example, it is better to take 2 strips with a volume of 2 Gb each, and put them in two-channel mode than with one 4 Gb strip.
Or 2 pieces of 1 Gb each, than one with a volume of 2 Gb.
The amount of memory is the same, but the performance gain is 10-70% more, depending on the application.
The only thing is that to provide a two-channel mode in an office computer, simple identical modules (preferably from one batch) are enough, then for home gaming, multimedia, gaming and overclocking systems, we strongly recommend purchasing KIT-memory (KIT-set).
The required amount of memory.
Today the minimum required amount of RAM is 2 Gb. ...
This is sufficient for any office system.
But the best choice is 4 Gb (2x2Gb). This is enough for any gaming machine.
Installation of 4 pieces is not desirable. 1 Gb each, this will lead to more power consumption and less stability when pairing in multi-channel mode.
Note: In order for the Windows operating system to use all 4 Gb of RAM, you need to install a 64-bit Windows operating system. Since a 32-bit system will use 3.12 Gb out of 4 Gb installed.
More RAM will be needed mainly by enthusiasts or professionals to process graphics and design models in high resolutions.
Installation of 8 Gb (2x4 Gb) and higher is justified in systems with SSD, and which uses a hard disk for short-term storage of files processed by RAM.
Disabling the paging file is only relevant on systems that use an SSD drive. To extend its service life.
And at the end of the article, I would like to say that there is never a lot of RAM, but there is no need for extra memory.
It is necessary to take exactly as much as necessary, and for the "extra" money choose Kit-memory with lower timings and with a higher frequency.
RAM is a special microcircuit used to store all kinds of data. There are many types of these devices, they are produced by a variety of companies. The best producers are most often of Japanese origin.
What is it and what is it for?
RAM (the so-called RAM memory) is a type of volatile microcircuit used to store all kinds of information. Most often it contains:
- machine code of programs currently executing (or in standby mode);
- input and output data.
Photo: RAM from different manufacturers
Data exchange between the central processor and RAM is carried out in two ways:
- using ultra-fast ALU register;
- through a special cache (if available in the design);
- directly (directly via the data bus).
The devices under consideration are circuits based on semiconductors. All information stored in all kinds of electronic components remains accessible only in the presence of an electric current. As soon as the voltage is turned off completely, or a short-term power cut occurs, then everything that was contained inside the RAM is erased or destroyed. ROM devices are an alternative.
Types and amount of memory
The board today can have a volume of several tens of gigabytes. Modern technical means allow you to use it as quickly as possible. Most operating systems are equipped with the ability to interact with such devices. There is a proportional relationship between the amount of RAM and the cost. The larger its size, the more expensive it is. And vice versa.
Also, the considered devices may have different frequencies. This parameter determines how quickly the interaction between RAM and other PC devices (CPU, data bus and video card) is carried out. The higher the operating speed, the more operations the PC will perform per unit of time.
The value of this characteristic also directly affects the cost of the device in question. The modern fastest modification can "memorize" 128 GB. It is produced by a company called Hynix and has the following performance characteristics:
All modern RAM can be divided into two types:
- static;
- dynamic.
Static type
More expensive today is the static microcircuit. It is labeled as SDRAM. Dynamic is cheaper.
The distinctive features of the SDRAM version are:
Also, a distinctive feature of RAM is the ability to select the bit in which any information will be recorded.
The disadvantages include:
- low recording density;
- relatively high cost.
Various types of computer random access memory devices (SDRAM and DRAM) have external differences. They are contained in the length of the contact part. Its shape also differs. The designation of the RAM is located both on the sticker label and printed directly on the strip itself.
There are many different modifications of SDRAM today. It is designated as:
- DDR 2;
- DDR 3;
- DDR 4.
Dynamic type
Another type of microcircuit is designated as DRAM. It is also completely volatile, with write bits being randomly accessed. This type is widely used in most modern PCs. It is also used in those computer systems where latency requirements are high - DRAM performance is an order of magnitude higher than SDRAM.
DRAM - dynamic memory
Most often, this type has a DIMM form factor. The same design solution is used for the manufacture of the static circuit (SDRAM). The peculiarity of the DIMM version is that there are contacts on both sides of the surface.
OP parameters
The main criteria for choosing microcircuits of this type are their operating parameters.
You should focus primarily on the following points:
- frequency of work;
- timings;
- voltage.
They all depend on the type of a particular model. For example, DDR 2 will perform various actions definitely faster than the DDR 1 bar. Since it has more outstanding performance characteristics.
Timings is the time delay of information between different components of the device. There are quite a few types of timings, all of them directly affect performance. Small timings allow you to increase the speed of various operations. There is one unpleasant proportional dependence - the higher the speed of the random access memory, the higher the timings.
The way out of this situation is to increase the operating voltage - the higher it is, the smaller the timings become. The number of operations performed per unit of time is increasing at the same time.
Frequency and speed
The higher the RAM bandwidth, the higher its speed. Frequency is a parameter that determines the bandwidth of the channels through which various kinds of data are transferred to the CPU through the motherboard.
It is desirable that this characteristic coincides with the permissible operating speed of the motherboard.
For example, if the bracket supports 1600 MHz, and the motherboard does not exceed 1066 MHz, then the speed of data exchange between the RAM and the CPU will be limited precisely by the capabilities of the motherboard. That is, the speed will be no more than 1066 MHz.
Performance
Performance depends on many factors. The number of planks used has a very large effect on this parameter. Dual-channel RAM works an order of magnitude faster than single-channel RAM. The presence of the ability to support multichannel modes is indicated on a sticker located over the board.
These designations are as follows:
To determine which mode is optimal for a particular motherboard, you need to count the total number of slots for connection and divide them by two. For example, if there are 4 of them, then you need 2 identical strips from the same manufacturer. When they are installed in parallel, the Dual mode is activated.
Working principle and functions
The operation of the OP is implemented quite simply, writing or reading data is carried out as follows:
Each column is connected to an extremely sensitive amplifier. It records the flow of electrons that occur when a capacitor is discharged. In this case, the corresponding command is given. Thus, access to the various cells located on the board takes place. There is one important nuance that you should definitely know. When an electrical impulse is applied to any line, it opens all its transistors. They are directly connected to it.
From this we can conclude that one line is the minimum amount of information that can be read when accessing. The main purpose of RAM is to store various kinds of temporary data that are necessary while the personal computer is turned on and the operating system is functioning. The most important executable files are loaded into RAM, the CPU executes them directly, simply saving the results of the operations performed.
Photo: interaction of memory with the processor
The cells also store:
- executable libraries;
- key codes, which were pressed;
- results of various mathematical operations.
If necessary, everything that is in RAM can be saved to the hard disk by the central processor. And to do it in the form in which it is necessary.
Manufacturers
In stores, you can find a huge amount of RAM from a variety of manufacturers. A large number of such products began to be supplied from Chinese companies.
To date, the most productive and high-quality products are the following brands:
- Kingston;
- Hynix;
- Corsair;
- Kingmax.
- Samsung.
It is a compromise between quality and performance.
RAM characteristics table
The same kind of random access memory from different manufacturers has similar performance characteristics.
That is why it is correct to carry out comparison, taking into account only the type:
Performance and price comparison
The performance of RAM directly depends on its cost. You can find out how much a DDR3 module costs in your nearest computer store, and you should also familiarize yourself with the price of DDR 1. Comparing their operating parameters and price, and then testing, you can easily be convinced of this.
It is most correct to compare RAM of the same type, but with different performance, depending on the operating frequency:
| Type of | Operating frequency, MHz | Cost, rub. | Speedwork, Aida 64,Memory Read, MB / s |
| DDR 3 | 1333 | 3190 | 19501 |
| DDR 3 | 1600 | 3590 | 22436 |
| DDR 3 | 1866 | 4134 | 26384 |
| DDR 3 | 2133 | 4570 | 30242 |
| DDR 3 | 2400 | 6548 | 33813 |
| DDR 3 | 2666 | 8234 | 31012 |
| DDR 3 | 2933 | 9550 | 28930 |
In Aida 64, all DDR 3s were tested on identical hardware:
- OS: Windows 8.1;
- CPU: i5-4670K;
- video card: GeForce GTX 780 Ti;
- motherboard: LGA1150, Intel Z87.
RAM is a very important part of a PC and greatly affects its performance. That is why, to increase it, it is recommended to set bars with a high frequency and small timings. This will give a big boost in computer performance, it is especially important for games and various professional programs.
Theoretical foundations and first results of low-level testing
DDR2 is a new memory standard approved by the Joint Electronic Device Engineering Council, which includes many manufacturers of microcircuits and memory modules, as well as chipsets. The early versions of the standard were published in March 2003, it was finally approved only in January 2004 and was named DDR2 SDRAM SPECIFICATION, JESD79-2, revision A (). DDR2 is based on the well-known and proven DDR (Double Data Rate) technology. You could even say, "DDR2 starts where DDR ends." In other words, the first DDR2s will operate at frequencies that are the limit for the current generation of DDR-400 memory (PC3200 standard, clock frequency 200 MHz), and its further variants will significantly surpass it. The first generation of DDR2 memory, already produced by such vendors as, and, are its varieties DDR2-400 and DDR2-533, operating at 200 MHz and 266 MHz, respectively. Further, a new generation of DDR2-667 and DDR2-800 modules is expected to appear, although it is noted that they are unlikely to appear at all and, moreover, will become widespread even by the end of this year.
For the sake of fairness, it should be noted that DDR2 memory, as such, appeared a long time ago - of course, we mean memory on video cards. Nevertheless, this type of DDR2 (called GDDR2) is actually a special type of memory designed specifically for the video card market and slightly different from the "desktop" DDR2, which is the subject of this review. general information
So, "desktop" DDR2-SDRAM is considered as an evolutionary replacement for the current memory generation - DDR. The principle of its operation is absolutely the same - data transfer (at the level of the memory module) is carried out via a 64-bit bus on both parts of the clock signal (upward - "edge", and downward - "cut"), which provides twice the effective data transfer rate in relation to its frequency. Of course, at the same time, DDR2 implements a number of innovations that allow a leap to much higher frequencies (and, therefore, higher bandwidth) and larger capacities of microcircuit arrays, on the one hand, and reduced power consumption of modules, on the other. How this is achieved, we will see later, but for now let us turn to the "macroscopic" facts. DDR2 memory modules are produced in a new form factor, in the form of 240-pin DIMM modules, which are electrically incompatible with slots for DDR memory modules (in terms of pin count, pin spacing and module pinout). Thus, the DDR2 standard is not backward compatible with DDR.
The table below shows the approved naming conventions and specifications for the first three DDR2 standards. It is easy to see that DDR2-400 has the same bandwidth as the current type of DDR-400 memory.
The first DDR2 memory modules will ship in 256MB, 512MB and 1GB variants. Nevertheless, the standard provides for the possibility of building modules of significantly larger capacity - up to 4 GB, which, however, are specialized modules (not compatible with the desktop versions, at least for the time being). In the future, the appearance of modules with even greater capacity is expected.
DDR2 chips will be manufactured using Fine Ball Grid Array (FBGA) packaging, which is more compact than the traditional TSOP-II, allowing for larger chip capacities with a smaller size and improved electrical and thermal performance. This packaging method is already used by some DDR manufacturers as an option, but it is recommended for use from the point of view of the JEDEC standard.
The voltage consumed by DDR2 modules, according to the standard, is 1.8 V, which is significantly lower than the supply voltage of DDR devices (2.5 V). A quite expected (although not so obvious) consequence of this fact is a decrease in power consumption, which is important for manufacturers, both laptops and large workstations and servers, where the problem of power dissipated by memory modules is far from the last place. DDR2 inside
The DDR2 standard includes several important data transfer related changes to the DDR specification that allow higher frequencies to be achieved with lower power consumption. How exactly the reduction in power dissipation is achieved while increasing the speed of the modules, we will look at right now.
Fetching data
The main change in DDR2 is the ability to sample 4 data bits per clock at once (4n-prefetch), as opposed to 2-bit sampling (2n-prefetch) implemented in DDR. In essence, this means that at each clock cycle of the memory bus, DDR2 transfers 4 bits of information from the logical (internal) banks of the memory chip to the I / O buffers along one data interface line, while ordinary DDR is capable of sending only 2 bits per clock per line. ... Quite naturally, the question arises - if this is so, then why is the effective bandwidth of DDR2-400 the same as that of a regular DDR-400 (3.2 GB / s), and not doubled?
To answer this question, let's first look at how ordinary DDR-400 memory works. In this case, both the memory core and the I / O buffers operate at 200 MHz, and the "effective" frequency of the external data bus, thanks to DDR technology, is 400 MHz. According to the 2n-prefetch rule, at each memory cycle (200 MHz), 2 bits of information are sent to the I / O buffer on each data interface line. The task of this buffer is to multiplex / demultiplex (MUX / DEMUX) the data stream - in a simple way, "distilling" a narrow high-speed stream into a wide low-speed stream, and vice versa. Since the logical banks in a DDR SDRAM memory chip have a data bus width between them and the level amplifier twice as wide as from the read latches to the external interface, the data buffer includes a type 2-1 multiplexer. In general, since memory chips, unlike modules, can have different data bus widths - usually x4 / x8 / x16 / x32, the use of such a MUX / DEMUX (2-1) scheme implemented in DDR means that the internal data stream with width X and transmission rate Y from the array is converted to an external stream with width X / 2 and frequency 2Y. This is called peak bandwidth balancing.
Let us now consider the operation scheme of a DDR2 SDRAM type memory microcircuit, equal-frequency and "equal-wide" (ie, the same data bus width) relative to the DDR-400 memory chip. First of all, we note that the width of the external data bus remains absolutely the same - 1 bit / line, as well as its effective frequency (in this example - 400 MHz). Actually, this is already enough to answer the question posed above - why the theoretical memory bandwidth of DDR2 and DDR memory modules are equal. Further, it is obvious that the use of a type 2-1 multiplexer used in DDR SDRAM is no longer suitable in the case of DDR2 SDRAM sampling data according to the 4n-prefetch rule. Instead, it requires the introduction of a more complex circuit with an additional conversion stage - a type 4-1 multiplexer. This means that the core output has become four times wider than the external interface of the microcircuit and by the same number of times lower in operating frequency. That is, by analogy with the above example, in the general case, the MUX / DEMUX 4-1 scheme converts the internal data stream with the width X and the transmission frequency Y from the array into the external stream with the width X / 4 and the frequency 4Y.
Since in this case the core of the memory microcircuits is synchronized at a frequency that is half that of the external one (100 MHz), while in DDR the internal and external data streams are synchronized at the same frequency (200 MHz), among the advantages of this approach is an increase in the percentage of useful chips and reduced energy consumption modules. By the way, this also explains why the DDR2 standard assumes the existence of memory modules with an "effective" frequency of 800 MHz - which is twice as high as in the current generation of DDR memory. After all, this "effective" DDR2 frequency can be achieved even now, with DDR-400 memory chips operating at their own frequency of 200 MHz, if we sample data according to the 4n-prefetch rule according to the scheme described above.
Thus, DDR2 means abandoning the extensive path of development of memory chips - in the sense of simply further increasing their frequency, which significantly complicates the production of stable working memory modules in large quantities. It is being replaced by an intensive development path associated with the expansion of the internal data bus (which is a mandatory and inevitable decision when using more complex multiplexing). We would venture to assume that in the future it is quite possible to expect the appearance of DDR4 memory, which fetch not 4, but immediately 8 bits of data from memory chips (according to the 8n-prefetch rule, using an 8-1 multiplexer), and working at a frequency not 2, but 4 times lower than the frequency of the I / O buffer :). Actually, there is nothing new in this approach - something like that has already been encountered in memory chips like Rambus DRAM. Nevertheless, it is easy to guess that the downside of this development path is the complication of the MUX / DEMUX device for the I / O buffer, which, in the case of DDR2, must serialize four data bits read in parallel. First of all, this should affect such an important characteristic of memory as its latency, which we will consider below.
Intrachip termination
The DDR2 standard also includes a number of other enhancements that improve various characteristics of the new memory type, including electrical ones. One of these innovations is in-chip signal termination. Its essence lies in the fact that to eliminate unnecessary electrical noise (due to signal reflection from the end of the line) on the memory bus, resistors are used to load the line not on the motherboard (as was the case with previous generations of memory), but inside the chips themselves. These resistors are deactivated when the chip is in operation and, conversely, are activated as soon as the chip enters the standby state. Since signal cancellation is now carried out much closer to its source, this eliminates electrical noise inside the memory chip during data transmission.
By the way, in connection with the technology of intra-chip termination, one cannot but dwell on such an issue as ... the heat dissipation of the module, which, in general, is primarily designed for the new DDR2 standard. Indeed, such a signal termination scheme leads to significant static currents inside the memory chips, which leads to their warming up. Well, this is indeed the case, although we note that the power consumed by the memory subsystem generally, from this it should not grow at all (it's just that the heat is now dissipating in another place). The problem here is a little different - namely, in the possibility of increasing the frequency of operation of such devices. It is very likely that this is why the first generation of DDR2 memory is not DDR2-800 at all, but only DDR2-400 and DDR2-533, for which the heat dissipation inside the chips is still at an acceptable level.
Additional delay
Additional latency (also known as "deferred CAS") is another enhancement introduced to the DDR2 standard to minimize the downtime of the command scheduler when transferring data from / to memory. To illustrate this (using reading as an example), let's start by reading Bank Interleave data from a DDR2 device with an additional delay of zero, which is equivalent to reading from a regular DDR memory.
At the first stage, the bank is opened using the ACTIVATE command, together with the submission of the first component of the address (the address of the line), which selects and activates the required bank and the line in its array. During the next cycle, the information is transferred to the internal data bus and routed to the level amplifier. When the amplified signal level reaches the required value (after a time has elapsed, called the delay between the determination of the row and column addresses, t RCD (RAS-to-CAS Delay), a READ with Auto-Precharge (RD_AP) command can be sent to execution in conjunction with column address to select the exact address of the data to be read from the level amplifier.After setting the read command, the column selection strobe is delayed - t CL (CAS signal delay, CAS Latency), during which the data selected from the level amplifier is synchronized and transmitted In this case, a situation may arise when the next command (ACTIVATE) cannot be sent for execution, since at this moment in time the execution of other commands has not yet ended. by one clock cycle, since at this moment the read command with automatic recharging (RD_AP) from bank 0 is already being executed. In addition, this leads to a break in the sequence of data output on the external bus, which reduces the real memory bandwidth.
To eliminate this situation and increase the efficiency of the command scheduler, DDR2 introduces the concept of an additional (additional) delay, t AL. When t AL is nonzero, the memory device monitors the READ (RD_AP) and WRITE (WR_AP) commands, but postpones their execution for a time equal to the amount of additional delay. Differences in the behavior of a DDR2 memory microcircuit with two different t AL values are shown in the figure.
The upper figure describes the operation mode of the DDR2 microcircuit at t AL = 0, which is equivalent to the operation of the DDR memory microcircuit device; the lower one corresponds to the case t AL = t RCD - 1, which is standard for DDR2. With this configuration, as can be seen from the figure, the ACTIVATE and READ commands can be executed one after the other. The actual implementation of the READ command will be delayed by the amount of additional delay, i.e. in fact, it will be executed at the same moment as in the diagram above.
The following figure shows an example of reading data from a DDR2 microcircuit under the assumption that t RCD = 4 clocks, which corresponds to t AL = 3 clocks. In this case, due to the introduction of additional latency, the ACTIVATE / RD_AP commands can be executed in a row, in turn, allowing data to be emitted in a continuous manner and maximizing the real memory bandwidth.
Delay in issuing CAS
As we saw above, DDR2, in terms of the external bus frequency, operates at higher speeds than DDR SDRAM. At the same time, since the new standard does not imply any significant changes in the technology of manufacturing the chips themselves, static delays at the DRAM device level should remain more or less constant. Typical intrinsic latency for DDR DRAM devices is 15 ns. For DDR-266 (with 7.5 ns cycle time) this is equivalent to two clocks, and for DDR2-533 (3.75 ns cycle time) - four.
As the memory frequencies increase further, it is necessary to multiply the number of supported CAS signal output latency values (towards b O higher values). The CAS latency values defined by the DDR2 standard are presented in the table. They are in the range of integers from 3 to 5 ticks; the use of fractional delays (multiples of 0.5) is not allowed in the new standard.
DRAM device latency is expressed in terms of cycle time (t CK), i.e. are equal to the product of the cycle time by the selected CAS delay value (t CL). Typical latency values for DDR2 devices fall within the 12-20 ns interval, on the basis of which the used CAS latency value is selected. Use b O higher latency values are impractical for reasons of memory subsystem performance, and lower ones - because of the need for stable operation of the memory device.
Delay recording
The DDR2 standard also changes the write latency specification (WRITE commands). The difference in the behavior of the write command in DDR and DDR2 devices is shown in the figure.
DDR SDRAM has a write latency of 1 clock. This means that the DRAM device begins to "capture" information on the data bus, on average, one clock cycle after the receipt of the WRITE command. However, given the increased speed of DDR2 devices, this period of time is too short for a DRAM device (namely, its I / O buffer) to successfully prepare for data capture. In this regard, the DDR2 standard defines the write latency as the delay in issuing CAS minus 1 clock cycle (t WL = t CL - 1). It is noted that tying the WRITE latency to the CAS latency not only allows you to achieve higher frequencies, but also simplifies the synchronization of read and write commands (setting Read-to-Write timings).
Recovery after recording
The procedure for writing to SDRAM memory is similar to a read operation with a difference in an additional interval t WR, which characterizes the recovery period of the interface after the operation (usually a push-pull delay between the end of data output to the bus and the initiation of a new cycle). This time interval, measured from the moment of the end of the write operation to the moment of entering the regeneration stage (Auto Precharge), ensures that the interface is restored after the write operation and guarantees the correctness of its execution. Note that the DDR2 standard does not change the specification of the write recovery period.
Thus, the latency of DDR2 devices as a whole can be considered one of the few characteristics by which the new standard loses to the DDR specifications. In this connection, it is quite obvious that the use of an equal-frequency DDR2 will hardly have any advantages in terms of speed over DDR. As it is in reality - as always, the results of the corresponding tests will show. Test results in RightMark Memory Analyzer
Well, now is the time to move on to the test results obtained in the test suite version 3.1. Let's remind that the main advantages of this test in relation to other available memory tests are wide functionality, openness of the methodology (the test is available to everyone for familiarization in the form) and thoroughly worked out documentation.
Testbed configurations and software
Test stand No. 1
- Processor: Intel Pentium 4 3.4 GHz (Prescott core, Socket 478, FSB 800 / HT, 1 MB L2) at 2.8 GHz
- Motherboard: ASUS P4C800 Deluxe on Intel 875P chipset
- Memory: 2x512 MB PC3200 DDR SDRAM DIMM TwinMOS (2.5-3-3-6 timings)
Test stand No. 2
- Processor: Intel Pentium 4 3.4 GHz (Prescott core, Socket 775, FSB 800 / HT, 1 MB L2) at 2.8 GHz
- Motherboard: Intel D915PCY on Intel 915 chipset
- Memory: 2x512 MB PC2-4300 DDR2 SDRAM DIMM Samsung (timings 4-4-4-8)
Software
- Windows XP Professional SP1
- Intel Chipset Installation Utility 5.0.2.1003
Maximum real memory bandwidth
The measurement of the maximum real memory bandwidth was carried out using a subtest Memory Bandwidth, presets Maximal RAM Bandwidth, Software Prefetch, MMX / SSE / SSE2... As the name of the selected presets suggests, this series of measurements uses the standard method for optimizing read operations from memory - Software Prefetch, the essence of which is to prefetch data that will be requested later from the main memory into the L2 cache of the processor. To optimize writing to memory, the Non-Temporal Store method is used to avoid clogging up the cache. The results using the MMX, SSE and SSE2 registers turned out to be almost identical - for example, below is a picture obtained on the Prescott / DDR2 platform using SSE2.
Prescott / DDR2, maximum real memory bandwidth
Note that there are no significant qualitative differences between DDR and DDR2 on equal-frequency Prescott in this test. But what is more interesting is that the quantitative characteristics of the memory bandwidth of DDR-400 and DDR2-533 turn out to be very close! (see table). And this is despite the fact that DDR2-533 memory has a maximum theoretical memory bandwidth of 8.6 GB / s (in dual-channel mode). Actually, we see nothing surprising in the result obtained - after all, the processor bus is still 800 MHz Quad-Pumped Bus, and its bandwidth is 6.4 GB / s, that's why it is the limiting factor.
As far as the efficiency of write operations in relation to reading is concerned, it is easy to see that it has remained the same. However, this again looks quite natural, since in this case the write bandwidth limit (2/3 of the read bandwidth) is clearly set by the microarchitectural features of the Prescott processor.
Memory latency
First of all, let's dwell in some detail on how and why we measured the "true" memory latency, since its measurement on Pentium 4 platforms is, in fact, a far from trivial task. This is due to the fact that the processors of this family, in particular, the new Prescott core, are characterized by the presence of a rather "advanced" asynchronous hardware data prefetcher, which makes it very difficult to objectively measure this characteristic of the memory subsystem. Obviously, the use of sequential memory bypass methods (forward or backward) to measure its latency is completely inappropriate in this case - the Hardware Prefetch algorithm in this case works with maximum efficiency, “masking” the latency. The use of random traversal modes is much more justified, however, true random traversal of memory has another significant drawback. The fact is that such a measurement is performed under conditions of almost 100% D-TLB miss, and this introduces significant additional delays, as we have already written about. Therefore, the only possible option (among the methods implemented in RMMA) is pseudo-random a memory traversal mode in which each subsequent page is loaded linearly (negating D-TLB misses), while a traversal within the memory page itself is truly random.
Nevertheless, the results of our past measurements have shown that even such a measurement technique underestimates the latency values quite strongly. We believe that this is due to another peculiarity of Pentium 4 processors, namely, the ability to "capture" two 64-byte lines from memory to L2 cache at once each time it is accessed. To demonstrate this phenomenon, the figure below shows the curves of the dependence of the latency of two consecutive accesses to the same memory line on the offset of the second line element relative to the first, obtained on the Prescott / DDR2 platform using the test D-Cache Arrival, preset L2 D-Cache Line Size Determination.
Prescott / DDR2, arrival of data on the L2-RAM bus
It can be seen from them (the random traversal curve is the most revealing) that access to the second line element is not accompanied by any additional delays up to 60 bytes inclusive (which corresponds to the true size of the L2 cache line, 64 bytes). Area 64-124 bytes corresponds to reading data from the next memory line. Since the latencies in this area increase only slightly, this means that the next line of memory is really "swapped" into the L2 cache of the processor immediately after the requested one. Which one can be made of all this practical output? The most direct: in order to "deceive" this feature of the Hardware Prefetch algorithm, which works in all memory bypass cases, it is enough to simply bypass the chain with a step equal to the so-called "effective" L2 cache line length, which in our case is 128 bytes.
So, let's go directly to the results of latency measurements. For clarity, here are the L2-RAM bus offload graphs obtained on the Prescott / DDR2 platform.
Prescott / DDR2, memory latency, 64 bytes line length
Prescott / DDR2, memory latency, 128 byte line length
As in the case of real memory bandwidth tests, the latency curves on another platform - Prescott / DDR - look exactly the same at a qualitative level. Only quantitative characteristics differ somewhat. It's time to turn to them.
* latency in the absence of offloading the L2-RAM bus
It is easy to see that the latency of DDR2-533 turned out to be higher than that of DDR-400. However, there is nothing supernatural here - according to the theoretical foundations of the new DDR2 memory standard presented above, this is how it should be.
The difference in latency between DDR and DDR2 is almost imperceptible with a standard 64-byte memory byte (3 ns in favor of DDR), when the hardware prefetcher is actively working, however, with a "two-line" (128-byte) chain walk it becomes much more noticeable. Namely, the minimum DDR2 latency (55.0 ns) is equal to the maximum DDR latency; if we compare the minimum and maximum latencies with each other, the difference is about 7-9 ns (15-16%) in favor of DDR. At the same time, I must say, the almost equal values of the "average" latency, obtained in the absence of unloading of the L2-RAM bus, are somewhat surprising - both in the case of 64-byte bypass (with data prefetch) and 128-byte (without such ). Conclusion
The main conclusion that suggests itself on the basis of the results of the first comparative testing of DDR and DDR2 memory, in general, can be formulated as follows: "The time for DDR2 has not come yet." The main reason is that it makes no sense to fight for an increase in the theoretical memory bandwidth by increasing the frequency of the external memory bus. After all, the bus of the current generation of processors still operates at 800 MHz, which limits the real bandwidth of the memory subsystem at 6.4 GB / s. This means that at present it makes no sense to install memory modules with a higher theoretical memory bandwidth, since the currently existing and widely used DDR-400 memory in dual-channel mode fully justifies itself, and in addition has a lower latency. Speaking of the latter, an increase in the frequency of the external memory bus is inevitably associated with the need to introduce additional delays, which, in fact, is confirmed by the results of our tests. Thus, we can assume that the use of DDR2 will justify itself, at least, not earlier than the moment when the first processors with a bus frequency of 1066 MHz and higher appear, which will allow overcoming the limitation imposed by the processor bus speed on the real bandwidth of the memory subsystem as a whole.
RAM frequency- the higher the frequency, the faster the information will be transferred for processing and the higher the performance of the computer. When talking about the frequency of RAM, they mean the frequency of data transmission, not the clock frequency.
- DDR- 200/266/333/400 MHz (clock frequency 100/133/166/200 MHz).
DDR2- 400/533/667/800/1066 MHz (200/266/333/400/533 MHz clock frequency). - DDR3- 800/1066/1333/1600/1800/2000/2133/2200/2400 MHz (400/533/667/800/1800/1000/1066/1100/1200 MHz clock frequency). But due to high timings (latencies), memory modules of the same frequency are inferior in performance to DDR2.
- DDR4 — 2133/2400/2666/2800/3000/3200/3333.
Data transmission frequency
Data transmission frequency (it is correct to call it - data transmission rate, Data rate) - the number of data transmission operations per second through the selected channel. Measured in gigatransfers (GT / s) or megatransfers (MT / s). For DDR3-1333, the baud rate will be 1333 MT / s.
You need to understand that this is not a clock frequency. The real frequency will be half of the indicated one, DDR (Double Data Rate) is the doubled data transfer rate. Therefore, DDR-400 memory operates at 200 MHz, DDR2-800 at 400 MHz, and DDR3-1333 at 666 MHz.
The frequency of the RAM indicated on the board is the maximum frequency with which it can work. If you install 2 boards DDR3-2400 and DDR3-1333, then the system will operate at the maximum frequency of the weakest board, i.e. at 1333. Thus, the throughput will decrease, but the decrease in throughput is not the only problem, errors may appear when loading the operating system and critical errors during operation. If you are going to buy RAM, you need to consider the frequency at which it can work. This frequency must match the frequency supported by the motherboard.
Maximum baud rate
The second parameter (in the photo PC3-10666) is the maximum data transfer rate measured in Mb / s. For DDR3-1333 PC3-10666, the maximum data transfer rate is 10.664 MB / s.
Timings and frequency of RAM
Many motherboards, when installing memory modules on them, do not set the maximum clock frequency for them. One of the reasons is the lack of performance gain when the clock frequency is increased, because when the frequency is increased, the operating timings increase. Of course, this can improve performance in some applications, but it can also decrease performance in others, or it may not affect applications that do not rely on memory latency or bandwidth at all.
Timing determines the memory delay time. For example, the CAS Latency parameter (CL, or access time) determines how many clock cycles of the memory module will delay the return of data requested by the processor. RAM with CL 9 will delay nine clock cycles to transfer the requested data, while RAM with CL 7 will delay seven clock cycles to transfer it. Both RAMs can have the same frequency and data transfer rates, but the second RAM will transfer data faster than the first. This problem is known as latency.
The smaller the timing parameter, the faster the memory.
For example. The Corsair memory module installed on the M4A79 Deluxe motherboard will have the following timings: 5-5-5-18. If you increase the clock frequency of the memory to DDR2-1066, the timings will increase and will have the following values 5-7-7-24.
The Qimonda memory module, when operating at a clock frequency of DDR3-1066, has operating timings of 7-7-7-20, when the operating frequency is increased to DDR3-1333, the board sets timings at 9-9-9-25. As a rule, the timings are written in the SPD and may differ for different modules.
