Dual-channel memory Dual-channel memory technology is actually a memory control and management technology. It relies on the memory controller of the chipset to act, and in theory can double the bandwidth provided by two memory sizes. It is not a new technology, it has long been applied to server and workstation systems. It has only reached the forefront of desktop motherboard technology to solve the increasingly difficult memory bandwidth bottleneck of desktop computers. A few years ago, Intel launched an i820 chipset that supports dual-channel memory transmission technology. It forms a golden partner with RDRAM memory. The outstanding performance it exerted makes it the biggest highlight of the market for a while, but the defect of excessive production costs has caused the popularity to be unpopular, and was eventually eliminated by the market. Since Intel has given up support for RDRAM, the current dual-channel memory technology of mainstream chipsets refers to dual-channel DDR memory technology. Intel is the Intel 865 and 875 series on the mainstream dual-channel memory platform, while AMD is the NVIDIA Nforce2 series.
Dual-channel memory technology is a low-cost and high-performance solution to solve the contradiction between CPU bus bandwidth and memory bandwidth. Now the CPU's FSB (front-end bus frequency) is getting higher and higher, and Intel Pentium 4 has much higher memory bandwidth requirements than AMD Athlon XP. The data transmission of Intel Pentium 4 processor and North Bridge chip uses QDR (Quad Data Rate, four data transmission) technology, and its FSB is 4 times that of external frequency. The FSBs of Intel Pentium 4 are 400, 533, and 800MHz, and the bus bandwidths are 3.2GB/sec, 4.2GB/sec and 6.4GB/sec, respectively. The memory bandwidths that DDR 266/DDR 333/DDR 400 can provide are 2.1GB/sec, 2.7GB/sec and 3.2GB/sec, respectively. In single-channel memory mode, DDR memory cannot provide the data bandwidth required by the CPU, thus becoming a performance bottleneck for the system. In dual-channel memory mode, the memory bandwidths that the dual-channel DDR 266, DDR 333, and DDR 400 can provide are 4.2GB/sec, 5.4GB/sec and 6.4GB/sec respectively. Here you can see that the dual-channel DDR 400 memory can just meet the bandwidth requirements of the 800MHz FSB Pentium 4 processor. For the AMD Athlon XP platform, the data transmission technology of its processor and North Bridge chip adopts DDR (Double Data Rate, double data transmission) technology, and the FSB is twice that of the external frequency. Its demand for memory bandwidth is much lower than that of Intel Pentium 4 platforms. Its FSB is 266, 333, and 400MHz, and its bus bandwidth is 2.1GB/sec, 2.7GB/sec and 3.2GB/sec, respectively. Using single-channel DDR 266, DDR 333, and DDR 400 can meet its bandwidth requirements. Therefore, using dual-channel DDR memory technology on the AMD K7 platform has little effect, and its performance improvement is not as obvious as that of Intel platform. The most obvious impact on performance is the integrated motherboard using integrated display chips.
NVIDIA's nForce chipset was the first to expand the DDR memory interface to 128-bit chipset. Intel then used this dual-channel DDR memory technology on its E7500 server motherboard chipset. SiS and VIA also responded and actively developed this technology that can double the DDR memory bandwidth. However, for various reasons, it is not easy for many chipset manufacturers to achieve this dual-channel DDR (128 bit parallel memory interface). DDR SDRAM memory is completely different from RDRAM memory. The latter has high latency and is a serial transmission method. These characteristics determine that the difficulty and cost of designing a dual-channel RDRAM memory chipset is not too high. However, DDR SDRAM memory has its own limitations. It is characterized by low latency and adopts parallel transmission mode. The most important point is: when the DDR SDRAM operates at a frequency of more than 400MHz, its signal waveform often has distortion problems. These all bring considerable difficulty to design a chipset that supports dual-channel DDR memory systems, and the manufacturing cost of the chipset will also increase accordingly. These factors restrict the development of this memory control technology.
A normal single-channel memory system has a 64-bit memory controller, while a dual-channel memory system has two 64-bit memory controllers, with a 128-bit memory bit width in dual-channel mode, which theoretically doubles the memory bandwidth. Although the bandwidth provided by a dual 64-bit memory system is equivalent to that provided by a 128-bit memory system, the results achieved by the two are different. The dual-channel system includes two independent, complementary intelligent memory controllers. In theory, both memory controllers can operate simultaneously with zero delay between each other. For example, two memory controllers, one is A and the other is B. When controller B is preparing to make the next access to memory, controller A is reading/writing the main memory, and vice versa. This complementary "nature" of the two memory controllers can reduce the waiting time by 50%. The two memory controllers of dual-channel DDR are exactly the same in function, and the timing parameters of both controllers can be programmed separately. This flexibility allows users to use two DIMM memory sticks with different structures, capacity and speeds. At this time, the dual-channel DDR is simply adjusted to the lowest memory standard to achieve 128 bit bandwidth, allowing DIMM memory sticks with different density/wait time characteristics to operate reliably.
Desktop chipsets that support dual-channel DDR memory technology, Intel platform includes Intel's 865P, 865G, 865GV, 865PE, 875P and the 915 and 925 series; VIA's PT880, ATI's Radeon 9100 IGP series, SIS SIIS 655, SIS 655FX and SIS 655TX; AMD platform includes VIA's KT880, NVIDIA's nForce2 Ultra 400, nForce2 IGP, nForce2 SPP and its future chips.
AMD's 64-bit CPU is integrated with a memory controller, so it is enough to see whether it supports dual-channel memory. Currently, AMD's desktop CPUs only support dual-channel memory for 939 interfaces, while 754 interfaces do not support dual-channel memory for 754 interfaces. In addition to AMD's 64-bit CPU, whether other computers can support dual-channel memory depends mainly on the motherboard chipset. The chipset that supports dual-channel is described above, and you can also view the motherboard chipset information. In addition, some chipsets theoretically support memory sticks with different capacity to implement dual channels, but in fact, it is recommended to use two memory sticks with the same parameters as much as possible.
The dual memory channels are generally required to be used in pairs according to the color of the memory slot on the motherboard. In addition, some motherboards also need to be set up in the BIOS, and the motherboard instruction manual will generally have instructions. When the system has realized dual channels, some motherboards will have prompts when powering on and self-test, so you can take a closer look. Because the self-test speed is relatively fast, it may not be visible. Therefore, you can use some software to view it, and many software can check it, such as cpu-z, which is relatively small. There is the "channels" item in the "memory" item. If the word "Dual" is displayed here, it means that dual channels have been implemented. The effect of two 256M memory forming a dual channel will be better than a 512M memory, because one memory cannot form a dual channel.
The difference between DDR and DDR2
Strictly speaking, DDR should be called DDR SDRAM, which people are used to being called DDR. Some beginners often see DDR SDRAM, which is considered SDRAM. DDR SDRAM is the abbreviation of Double Data Rate SDRAM, which means double-rate synchronous dynamic random memory. DDR memory is developed based on SDRAM memory and still uses the SDRAM production system. Therefore, for memory manufacturers, just a slight improvement on the equipment that manufactures ordinary SDRAM can realize the production of DDR memory, which can effectively reduce costs.
SDRAM only transmits data once in a clock cycle, and it transmits data during the rising period of the clock; while DDR memory transmits data twice in a clock cycle, and it can transmit data once in each of the rising period and falling period of the clock, so it is called double-rate synchronous dynamic random memory. DDR memory can achieve higher data transmission rates at the same bus frequency as SDRAM.
Compared with SDRAM: DDR uses more advanced synchronization circuits to enable the main steps of the transmission and output of specified addresses and data to be performed independently and remain completely synchronized with the CPU; DDR uses DLL (Delay Locked Loop, the delay locking loop provides a data filtering signal) technology. When the data is valid, the storage controller can use this data filtering signal to accurately locate the data, output it every 16 times, and resynchronize the data from different memory modules. DDL essentially doubles the speed of the SDRAM without increasing the clock frequency, which allows data to be read out on the rising and falling edges of the clock pulse, thus twice as fast as standard SDRA.
DDRs are not much different from SDRAM in terms of appearance volume, they have the same size and the same pin distance. However, DDR is 184 pins, which has 16 more pins than SDRAM, mainly including new control, clock, power supply and ground signals. DDR memory uses the SSTL2 standard that supports 2.5V voltage, rather than the LVTTL standard that uses 3.3V voltage, used by SDRAM.
The DDR2 memory starts from the highest standard frequency of DDR memory 400Mhz. The frequency that can be produced has been defined to support 533Mhz to 667Mhz. The standard operating frequency is 200/266/333MHz, and the operating voltage is 1.8V. DDR2 adopts the newly defined 240 PIN DIMM interface standard, which is completely incompatible with DDR's 184 PIN DIMM interface standard.
Like DDR, DDR2 adopts the basic method of data transmission at the same time as the clock rise and fall delays, but the biggest difference is that DDR2 memory can perform 4-bit pre-reading. 2BIT pre-reading twice as high as standard DDR memory means that DDR2 has twice as much ability to read pre-read system command data from DDR, so DDR2 simply obtains twice as much complete data transmission capability as DDR.
The biggest breakthrough point of DDR2 memory technology is actually not the so-called transmission capacity that is twice that of DDR. Instead, when using lower heat generation and lower power consumption, it will achieve faster frequency improvement, breaking through the 400MHZ limit of standard DDR.
The difference between DDR2 and DDR
Compared with DDR, the most important improvement of DDR2 is that it can provide bandwidth equivalent to twice the DDR memory when the memory module speed is the same. This is mainly achieved by using two DRAM cores efficiently on each device. For comparison, DDR memory can only use one DRAM core on each device. Technically, there is still only one DRAM core on DDR2 memory, but it can be accessed in parallel, processing 4 data instead of two data in each access.
Schematic diagram of the difference between DDR2 and DDR
Combined with double-speed data buffering, DDR2 memory enables processing up to 4 bits of data per clock cycle, which is twice as high as 2 bits of data that traditional DDR memory can handle. Another improvement in DDR2 memory is that it uses FBGA packaging to replace the traditional TSOP method.
However, although the DRAM core speed used by DDR2 memory is the same as that of DDR, we still have to use the new motherboard to match DDR2 memory, because the physical specifications of DDR2 are incompatible with DDR. First of all, the interfaces are different. The number of pins of DDR2 is 240 pins, while the DDR memory is 184 pins; secondly, the VDIMM voltage of DDR2 memory is 1.8V, which is also different from the 2.5V of DDR memory.
Definition of DDR2:
DDR2 (Double Data Rate 2) SDRAM is a new generation memory technology standard developed by JEDEC (Joint Committee on Electronic Equipment Engineering). The biggest difference between it and the previous generation DDR memory technology standard is that although it adopts the basic method of data transmission at the same time during the rising/descending delay of the clock, DDR2 memory has twice the capability of pre-reading the previous generation DDR memory (i.e., 4bit data read prefetch). In other words, each clock in DDR2 memory can read/write data at 4 times the speed of the external bus and can run at 4 times the speed of the internal control bus.
In addition, since the DDR2 standard stipulates that all DDR2 memory is in the FBGA package form, unlike the TSOP/TSOP-II package form that is widely used, the FBGA package can provide better electrical performance and heat dissipation, providing a solid foundation for the stable operation of DDR2 memory and the future frequency development. Looking back on the development history of DDR, from the first generation of DDR200 applied to personal computers, through DDR266 and DDR333 to today's dual-channel DDR400 technology, the development of the first generation of DDR has also reached the limit of technology, and it is difficult to improve the working speed of memory through conventional methods; with the development of Intel's latest processor technology, the requirements for memory bandwidth of the front-end bus are getting higher and higher, and DDR2 memory with higher and more stable operating frequency will be the general trend.
The difference between DDR2 and DDR:
Before understanding many new DDR2 memory technologies, let’s first look at a set of data comparing DDR and DDR2 technologies.
1. Delay problem:
As can be seen from the above table, at the same core frequency, the actual operating frequency of DDR2 is twice that of DDR. This is due to the fact that DDR2 memory has twice the 4BIT pre-read capability of standard DDR memory. In other words, although DDR2, like DDR, adopts the basic method of data transmission at the same time as the clock rise and fall delays, DDR2 has the ability to pre-read system command data twice as much as DDR. That is to say, at the same operating frequency of 100MHz, the actual frequency of DDR is 200MHz, while DDR2 can reach 400MHz.
This also leads to another problem: in DDR and DDR2 memory with the same operating frequency, the memory delay of the latter is slower than the former. For example, DDR 200 and DDR2-400 have the same latency, while the latter has twice the bandwidth. In fact, DDR2-400 and DDR 400 have the same bandwidth, both of which are 3.2GB/s, but the core operating frequency of DDR400 is 200MHz, while the core operating frequency of DDR2-400 is 100MHz, which means that the latency of DDR2-400 is higher than that of DDR400.
2. Packaging and heating capacity:
The biggest breakthrough point of DDR2 memory technology is actually not what users think is twice the transmission capacity of DDR, but that DDR2 can obtain faster frequency improvements when using lower heat generation and lower power consumption, breaking through the 400MHZ limit of standard DDR.
DDR memory is usually in the form of TSOP chip package, which can work well at 200MHz. When the frequency is higher, its too long pins will produce high impedance and parasitic capacitance, which will affect its stability and the difficulty of increasing frequency. This is why it is difficult for DDR to break through 275MHZ. DDR2 memory is all in FBGA encapsulation. Unlike the TSOP packaging form that is widely used at present, the FBGA packaging provides better electrical performance and heat dissipation, providing good guarantees for the stable operation of DDR2 memory and the development of future frequency.
The DDR2 memory uses a 1.8V voltage, which is a lot lower than the DDR standard 2.5V, thereby providing significantly less power consumption and less heat generation. This change is of great significance.
New technologies adopted by DDR2:
In addition to the differences mentioned above, DDR2 also introduces three new technologies, namely OCD, ODT and Post CAS.
OCD (Off-Chip Driver): This is the so-called offline driver adjustment. DDR II can improve signal integrity through OCD. DDR II adjusts the resistance value of pull-up/pull-down to make the voltages of the two equal. The use of OCD improves signal integrity by reducing the tilt of DQ-DQS; and improves signal quality by controlling the voltage.
ODT: ODT is a termination resistor with built-in core. We know that a large number of termination resistors are required on the motherboard using DDR SDRAM to prevent the data line terminal from reflecting signals. It greatly increases the manufacturing cost of motherboards. In fact, different memory modules have different requirements for the termination circuit. The magnitude of the termination resistor determines the signal ratio and reflectivity of the data line. If the termination resistor is small, the signal-to-noise ratio of the data line will be low, but the signal-to-noise ratio will be low; if the termination resistor is high, the signal-to-noise ratio of the data line will be high, but the signal-to-noise ratio will also increase. Therefore, the termination resistor on the motherboard cannot match the memory module very well, and will also affect the signal quality to a certain extent. DDR2 can build a suitable termination resistor according to its own characteristics, which can ensure the optimal signal waveform. Using DDR2 not only reduces the motherboard cost, but also obtains the best signal quality, which is incomparable to DDR.
Post CAS: It is set to improve the utilization efficiency of DDR II memory. In Post CAS operation, the CAS signal (read and write/command) can be inserted into a clock cycle after the RAS signal, and the CAS command can remain valid after an additional delay (Additive Latency). The original tRCD (RAS to CAS and delay) is replaced by AL (Additive Latency), which can be set in 0, 1, 2, 3, 4. Since the CAS signal is placed one clock cycle after the RAS signal, the ACT and CAS signals will never have collision conflicts.
In general, DDR2 adopts many new technologies to improve many shortcomings of DDR. Although it currently has many shortcomings in cost and slow delay, I believe that with the continuous improvement and improvement of technology, these problems will eventually be solved.
Dual-channel memory technology is a low-cost and high-performance solution to solve the contradiction between CPU bus bandwidth and memory bandwidth. Now the CPU's FSB (front-end bus frequency) is getting higher and higher, and Intel Pentium 4 has much higher memory bandwidth requirements than AMD Athlon XP. The data transmission of Intel Pentium 4 processor and North Bridge chip uses QDR (Quad Data Rate, four data transmission) technology, and its FSB is 4 times that of external frequency. The FSBs of Intel Pentium 4 are 400, 533, and 800MHz, and the bus bandwidths are 3.2GB/sec, 4.2GB/sec and 6.4GB/sec, respectively. The memory bandwidths that DDR 266/DDR 333/DDR 400 can provide are 2.1GB/sec, 2.7GB/sec and 3.2GB/sec, respectively. In single-channel memory mode, DDR memory cannot provide the data bandwidth required by the CPU, thus becoming a performance bottleneck for the system. In dual-channel memory mode, the memory bandwidths that the dual-channel DDR 266, DDR 333, and DDR 400 can provide are 4.2GB/sec, 5.4GB/sec and 6.4GB/sec respectively. Here you can see that the dual-channel DDR 400 memory can just meet the bandwidth requirements of the 800MHz FSB Pentium 4 processor. For the AMD Athlon XP platform, the data transmission technology of its processor and North Bridge chip adopts DDR (Double Data Rate, double data transmission) technology, and the FSB is twice that of the external frequency. Its demand for memory bandwidth is much lower than that of Intel Pentium 4 platforms. Its FSB is 266, 333, and 400MHz, and its bus bandwidth is 2.1GB/sec, 2.7GB/sec and 3.2GB/sec, respectively. Using single-channel DDR 266, DDR 333, and DDR 400 can meet its bandwidth requirements. Therefore, using dual-channel DDR memory technology on the AMD K7 platform has little effect, and its performance improvement is not as obvious as that of Intel platform. The most obvious impact on performance is the integrated motherboard using integrated display chips.
NVIDIA's nForce chipset was the first to expand the DDR memory interface to 128-bit chipset. Intel then used this dual-channel DDR memory technology on its E7500 server motherboard chipset. SiS and VIA also responded and actively developed this technology that can double the DDR memory bandwidth. However, for various reasons, it is not easy for many chipset manufacturers to achieve this dual-channel DDR (128 bit parallel memory interface). DDR SDRAM memory is completely different from RDRAM memory. The latter has high latency and is a serial transmission method. These characteristics determine that the difficulty and cost of designing a dual-channel RDRAM memory chipset is not too high. However, DDR SDRAM memory has its own limitations. It is characterized by low latency and adopts parallel transmission mode. The most important point is: when the DDR SDRAM operates at a frequency of more than 400MHz, its signal waveform often has distortion problems. These all bring considerable difficulty to design a chipset that supports dual-channel DDR memory systems, and the manufacturing cost of the chipset will also increase accordingly. These factors restrict the development of this memory control technology.
A normal single-channel memory system has a 64-bit memory controller, while a dual-channel memory system has two 64-bit memory controllers, with a 128-bit memory bit width in dual-channel mode, which theoretically doubles the memory bandwidth. Although the bandwidth provided by a dual 64-bit memory system is equivalent to that provided by a 128-bit memory system, the results achieved by the two are different. The dual-channel system includes two independent, complementary intelligent memory controllers. In theory, both memory controllers can operate simultaneously with zero delay between each other. For example, two memory controllers, one is A and the other is B. When controller B is preparing to make the next access to memory, controller A is reading/writing the main memory, and vice versa. This complementary "nature" of the two memory controllers can reduce the waiting time by 50%. The two memory controllers of dual-channel DDR are exactly the same in function, and the timing parameters of both controllers can be programmed separately. This flexibility allows users to use two DIMM memory sticks with different structures, capacity and speeds. At this time, the dual-channel DDR is simply adjusted to the lowest memory standard to achieve 128 bit bandwidth, allowing DIMM memory sticks with different density/wait time characteristics to operate reliably.
Desktop chipsets that support dual-channel DDR memory technology, Intel platform includes Intel's 865P, 865G, 865GV, 865PE, 875P and the 915 and 925 series; VIA's PT880, ATI's Radeon 9100 IGP series, SIS SIIS 655, SIS 655FX and SIS 655TX; AMD platform includes VIA's KT880, NVIDIA's nForce2 Ultra 400, nForce2 IGP, nForce2 SPP and its future chips.
AMD's 64-bit CPU is integrated with a memory controller, so it is enough to see whether it supports dual-channel memory. Currently, AMD's desktop CPUs only support dual-channel memory for 939 interfaces, while 754 interfaces do not support dual-channel memory for 754 interfaces. In addition to AMD's 64-bit CPU, whether other computers can support dual-channel memory depends mainly on the motherboard chipset. The chipset that supports dual-channel is described above, and you can also view the motherboard chipset information. In addition, some chipsets theoretically support memory sticks with different capacity to implement dual channels, but in fact, it is recommended to use two memory sticks with the same parameters as much as possible.
The dual memory channels are generally required to be used in pairs according to the color of the memory slot on the motherboard. In addition, some motherboards also need to be set up in the BIOS, and the motherboard instruction manual will generally have instructions. When the system has realized dual channels, some motherboards will have prompts when powering on and self-test, so you can take a closer look. Because the self-test speed is relatively fast, it may not be visible. Therefore, you can use some software to view it, and many software can check it, such as cpu-z, which is relatively small. There is the "channels" item in the "memory" item. If the word "Dual" is displayed here, it means that dual channels have been implemented. The effect of two 256M memory forming a dual channel will be better than a 512M memory, because one memory cannot form a dual channel.
The difference between DDR and DDR2
Strictly speaking, DDR should be called DDR SDRAM, which people are used to being called DDR. Some beginners often see DDR SDRAM, which is considered SDRAM. DDR SDRAM is the abbreviation of Double Data Rate SDRAM, which means double-rate synchronous dynamic random memory. DDR memory is developed based on SDRAM memory and still uses the SDRAM production system. Therefore, for memory manufacturers, just a slight improvement on the equipment that manufactures ordinary SDRAM can realize the production of DDR memory, which can effectively reduce costs.
SDRAM only transmits data once in a clock cycle, and it transmits data during the rising period of the clock; while DDR memory transmits data twice in a clock cycle, and it can transmit data once in each of the rising period and falling period of the clock, so it is called double-rate synchronous dynamic random memory. DDR memory can achieve higher data transmission rates at the same bus frequency as SDRAM.
Compared with SDRAM: DDR uses more advanced synchronization circuits to enable the main steps of the transmission and output of specified addresses and data to be performed independently and remain completely synchronized with the CPU; DDR uses DLL (Delay Locked Loop, the delay locking loop provides a data filtering signal) technology. When the data is valid, the storage controller can use this data filtering signal to accurately locate the data, output it every 16 times, and resynchronize the data from different memory modules. DDL essentially doubles the speed of the SDRAM without increasing the clock frequency, which allows data to be read out on the rising and falling edges of the clock pulse, thus twice as fast as standard SDRA.
DDRs are not much different from SDRAM in terms of appearance volume, they have the same size and the same pin distance. However, DDR is 184 pins, which has 16 more pins than SDRAM, mainly including new control, clock, power supply and ground signals. DDR memory uses the SSTL2 standard that supports 2.5V voltage, rather than the LVTTL standard that uses 3.3V voltage, used by SDRAM.
The DDR2 memory starts from the highest standard frequency of DDR memory 400Mhz. The frequency that can be produced has been defined to support 533Mhz to 667Mhz. The standard operating frequency is 200/266/333MHz, and the operating voltage is 1.8V. DDR2 adopts the newly defined 240 PIN DIMM interface standard, which is completely incompatible with DDR's 184 PIN DIMM interface standard.
Like DDR, DDR2 adopts the basic method of data transmission at the same time as the clock rise and fall delays, but the biggest difference is that DDR2 memory can perform 4-bit pre-reading. 2BIT pre-reading twice as high as standard DDR memory means that DDR2 has twice as much ability to read pre-read system command data from DDR, so DDR2 simply obtains twice as much complete data transmission capability as DDR.
The biggest breakthrough point of DDR2 memory technology is actually not the so-called transmission capacity that is twice that of DDR. Instead, when using lower heat generation and lower power consumption, it will achieve faster frequency improvement, breaking through the 400MHZ limit of standard DDR.
The difference between DDR2 and DDR
Compared with DDR, the most important improvement of DDR2 is that it can provide bandwidth equivalent to twice the DDR memory when the memory module speed is the same. This is mainly achieved by using two DRAM cores efficiently on each device. For comparison, DDR memory can only use one DRAM core on each device. Technically, there is still only one DRAM core on DDR2 memory, but it can be accessed in parallel, processing 4 data instead of two data in each access.
Schematic diagram of the difference between DDR2 and DDR
Combined with double-speed data buffering, DDR2 memory enables processing up to 4 bits of data per clock cycle, which is twice as high as 2 bits of data that traditional DDR memory can handle. Another improvement in DDR2 memory is that it uses FBGA packaging to replace the traditional TSOP method.
However, although the DRAM core speed used by DDR2 memory is the same as that of DDR, we still have to use the new motherboard to match DDR2 memory, because the physical specifications of DDR2 are incompatible with DDR. First of all, the interfaces are different. The number of pins of DDR2 is 240 pins, while the DDR memory is 184 pins; secondly, the VDIMM voltage of DDR2 memory is 1.8V, which is also different from the 2.5V of DDR memory.
Definition of DDR2:
DDR2 (Double Data Rate 2) SDRAM is a new generation memory technology standard developed by JEDEC (Joint Committee on Electronic Equipment Engineering). The biggest difference between it and the previous generation DDR memory technology standard is that although it adopts the basic method of data transmission at the same time during the rising/descending delay of the clock, DDR2 memory has twice the capability of pre-reading the previous generation DDR memory (i.e., 4bit data read prefetch). In other words, each clock in DDR2 memory can read/write data at 4 times the speed of the external bus and can run at 4 times the speed of the internal control bus.
In addition, since the DDR2 standard stipulates that all DDR2 memory is in the FBGA package form, unlike the TSOP/TSOP-II package form that is widely used, the FBGA package can provide better electrical performance and heat dissipation, providing a solid foundation for the stable operation of DDR2 memory and the future frequency development. Looking back on the development history of DDR, from the first generation of DDR200 applied to personal computers, through DDR266 and DDR333 to today's dual-channel DDR400 technology, the development of the first generation of DDR has also reached the limit of technology, and it is difficult to improve the working speed of memory through conventional methods; with the development of Intel's latest processor technology, the requirements for memory bandwidth of the front-end bus are getting higher and higher, and DDR2 memory with higher and more stable operating frequency will be the general trend.
The difference between DDR2 and DDR:
Before understanding many new DDR2 memory technologies, let’s first look at a set of data comparing DDR and DDR2 technologies.
1. Delay problem:
As can be seen from the above table, at the same core frequency, the actual operating frequency of DDR2 is twice that of DDR. This is due to the fact that DDR2 memory has twice the 4BIT pre-read capability of standard DDR memory. In other words, although DDR2, like DDR, adopts the basic method of data transmission at the same time as the clock rise and fall delays, DDR2 has the ability to pre-read system command data twice as much as DDR. That is to say, at the same operating frequency of 100MHz, the actual frequency of DDR is 200MHz, while DDR2 can reach 400MHz.
This also leads to another problem: in DDR and DDR2 memory with the same operating frequency, the memory delay of the latter is slower than the former. For example, DDR 200 and DDR2-400 have the same latency, while the latter has twice the bandwidth. In fact, DDR2-400 and DDR 400 have the same bandwidth, both of which are 3.2GB/s, but the core operating frequency of DDR400 is 200MHz, while the core operating frequency of DDR2-400 is 100MHz, which means that the latency of DDR2-400 is higher than that of DDR400.
2. Packaging and heating capacity:
The biggest breakthrough point of DDR2 memory technology is actually not what users think is twice the transmission capacity of DDR, but that DDR2 can obtain faster frequency improvements when using lower heat generation and lower power consumption, breaking through the 400MHZ limit of standard DDR.
DDR memory is usually in the form of TSOP chip package, which can work well at 200MHz. When the frequency is higher, its too long pins will produce high impedance and parasitic capacitance, which will affect its stability and the difficulty of increasing frequency. This is why it is difficult for DDR to break through 275MHZ. DDR2 memory is all in FBGA encapsulation. Unlike the TSOP packaging form that is widely used at present, the FBGA packaging provides better electrical performance and heat dissipation, providing good guarantees for the stable operation of DDR2 memory and the development of future frequency.
The DDR2 memory uses a 1.8V voltage, which is a lot lower than the DDR standard 2.5V, thereby providing significantly less power consumption and less heat generation. This change is of great significance.
New technologies adopted by DDR2:
In addition to the differences mentioned above, DDR2 also introduces three new technologies, namely OCD, ODT and Post CAS.
OCD (Off-Chip Driver): This is the so-called offline driver adjustment. DDR II can improve signal integrity through OCD. DDR II adjusts the resistance value of pull-up/pull-down to make the voltages of the two equal. The use of OCD improves signal integrity by reducing the tilt of DQ-DQS; and improves signal quality by controlling the voltage.
ODT: ODT is a termination resistor with built-in core. We know that a large number of termination resistors are required on the motherboard using DDR SDRAM to prevent the data line terminal from reflecting signals. It greatly increases the manufacturing cost of motherboards. In fact, different memory modules have different requirements for the termination circuit. The magnitude of the termination resistor determines the signal ratio and reflectivity of the data line. If the termination resistor is small, the signal-to-noise ratio of the data line will be low, but the signal-to-noise ratio will be low; if the termination resistor is high, the signal-to-noise ratio of the data line will be high, but the signal-to-noise ratio will also increase. Therefore, the termination resistor on the motherboard cannot match the memory module very well, and will also affect the signal quality to a certain extent. DDR2 can build a suitable termination resistor according to its own characteristics, which can ensure the optimal signal waveform. Using DDR2 not only reduces the motherboard cost, but also obtains the best signal quality, which is incomparable to DDR.
Post CAS: It is set to improve the utilization efficiency of DDR II memory. In Post CAS operation, the CAS signal (read and write/command) can be inserted into a clock cycle after the RAS signal, and the CAS command can remain valid after an additional delay (Additive Latency). The original tRCD (RAS to CAS and delay) is replaced by AL (Additive Latency), which can be set in 0, 1, 2, 3, 4. Since the CAS signal is placed one clock cycle after the RAS signal, the ACT and CAS signals will never have collision conflicts.
In general, DDR2 adopts many new technologies to improve many shortcomings of DDR. Although it currently has many shortcomings in cost and slow delay, I believe that with the continuous improvement and improvement of technology, these problems will eventually be solved.