What Is an SDRAM Card?
SDRAM is the abbreviation of Synchronous Dynamic Random Access Memory. SDRAM uses a 3.3v operating voltage and a bandwidth of 64 bits. SDRAM locks the CPU and RAM together with the same clock so that the RAM and the CPU can share a clock cycle. , Synchronous work at the same speed, 50% faster than EDO memory. SDRAM is based on a dual-bank structure and contains two interleaved memory arrays. When the CPU accesses data from one bank or array, the other is ready for reading and writing data. Tight switching between the two memory array , The reading efficiency can be doubled. SDRAM is not only used as main memory, it is also widely used in graphics card memory.
SDRAM memory
- SDRAM used to be the mainstream for a long time
- The difference between DDR2 and DDR
- Compared with DDR, the main improvement of DDR2 is in
- Since the launch of the Intel Celeron series and AMD K6 processors and related motherboard chipsets, the performance of EDO DRAM memory can no longer meet the needs. The memory technology must be completely reformed to meet the needs of the new generation of CPU architecture. At this time, the memory begins to enter More classic SDRAM era.
- The first generation of SDRAM memory was the PC66 specification, but soon the CPU FSB was increased to 100MHz due to the frequency dispute between Intel and AMD, so the PC66 memory was quickly replaced by the PC100 memory, and then the PIII of 133MHz FSB and the advent of the K7 era The PC133 specification also further improves the overall performance of SDRAM in the same way, and the bandwidth is increased to more than 1GB / sec. Because the bandwidth of SDRAM is 64bit, which corresponds to the 64bit data bus width of the CPU, it only needs one memory to work, and the convenience is further improved. In terms of performance, because its input and output signals remain synchronized with the system's external frequency, the speed is significantly faster than EDO memory.
- It is undeniable that SDRAM memory developed from the early 66MHz to the later 100MHz and 133MHz. Although the bottleneck of memory bandwidth was not completely solved, at this time, CPU overclocking has become an eternal topic for DIY users. PC100 brand memory is overclocked to 133MHz for CPU overclocking success. It is worth mentioning that in order to facilitate the needs of some overclocking users, some PC150 and PC166 specifications have appeared on the market.
- Although the bandwidth of SDRAM PC133 memory can increase the bandwidth to 1064MB / S, and Intel has already begun the latest Pentium 4 plan, so SDRAM PC133 memory cannot meet future development needs. At this time, in order to achieve the purpose of exclusive market, Intel and Rambus Jointly promoted Rambus DRAM memory (called RDRAM memory) in the PC market. Different from SDRAM, it uses a new generation of high-speed and simple memory architecture, based on a RISC (Reduced Instruction Set Computing). This theory can reduce the complexity of data and improve the performance of the entire system.
- Nonetheless, Rambus RDRAM memory is not in a good time, and it will still be "robbed" of its throne position by higher-speed DDR. At that time, the Ram600 RDRAM memory of PC600 and PC700 appeared due to Intel820 chipset "error event" and PC800 Rambus RDRAM. Due to the high cost, the Pentium 4 platform was high and could not gain the support of the masses. Rambus RDRAM was killed. Rambus had hoped that the higher frequency PC1066 standard RDRAM could turn the tide, but it finally fell over in front of DDR memory.
- 1.DDR era
- DDR SDRAM (Double Data Rate SDRAM) is referred to as DDR, which means "double rate SDRAM". DDR can be said to be an upgraded version of SDRAM. DDR transmits data once at the rising and falling edges of the clock signal, which makes the data transmission speed of DDR twice that of traditional SDRAM. Since only more falling edge signals are used, no increase in energy consumption is caused. As for the addressing and control signals, they are the same as traditional SDRAM, and are transmitted only on the rising edge of the clock.
- DDR memory is a solution that is a compromise between performance and cost. The purpose is to quickly establish a solid market space, then step by step in frequency, and finally make up for the lack of memory bandwidth. The first generation of DDR200 specifications did not gain popularity. The second generation of PC266 DDR SRAM (133MHz clock x 2 times data transmission = 266MHz bandwidth) was derived from PC133 SDRAM memory. It brought DDR memory to the first climax. Currently, There are also many Celeron and AMD K7 processors that are using DDR266 memory. Later DDR333 memory is also a kind of transition, and DDR400 memory has become the mainstream platform option. Dual-channel DDR400 memory has become 800FSB With the basic standard, the subsequent DDR533 specification has become the object of choice for overclocking users.
- 2.DDR2 era
- As CPU performance continues to improve, our requirements for memory performance have also gradually increased. It is undeniable that DDR, which depends on the high frequency to increase the bandwidth, will sooner or later fail, so the JEDEC organization began to brew the DDR2 standard very early. In addition, 915/925 with the LGA775 interface and the latest 945 and other new platforms began to support DDR2 memory. DDR2 memory will begin to define the memory field today.
- DDR2 can provide a minimum bandwidth of 400MB / s per pin based on the 100MHz transmission frequency, and its interface will run at 1.8V, which will further reduce the heat generation and increase the frequency. In addition, DDR2 will incorporate new performance indicators and interrupt instructions such as CAS, OCD, ODT, and improve the utilization of memory bandwidth. Judging from the DDR2 standard described by the JEDEC organizers, DDR2 memory for the PC and other markets will have different clock frequencies such as 400, 533, and 667MHz. High-end DDR2 memory will have two frequencies of 800 and 1000MHz. DDR-II memory will be available in 200-, 220-, and 240-pin FBGA packages. The initial DDR2 memory will use a 0.13 micron production process, the voltage of the memory particles is 1.8V, and the capacity density is 512MB.
- Memory technology will have no suspense in 2005, and static memory represented by SDRAM will not be popular in five years. QBM and RDRAM memory are also difficult to recover, so the era of coexistence of DDR and DDR2 will be a certainty.
- In addition to PC-133, the "successor" of PC-100, VCM (VirXual Channel Memory) is also an important member. VCM is "Virtual Channel Memory", which is also a memory standard supported by most recent newer chipset. VCM memory is mainly manufactured based on a "cached DRAM" technology developed by NEC Corporation, which integrates " Channel buffer "is configured and controlled by high-speed registers. While achieving high-speed data transmission, VCM also maintains a high degree of compatibility with traditional SDRAM, so VCM memory is often referred to as VCM SDRAM. The difference between VCM and SDRAM is that regardless of whether the data is processed by the CPU, it can be submitted to the VCM for processing, while ordinary SDRAM can only process the data processed by the CPU, so VCM can process data 20% faster than SDRAM. the above. There are many chipsets that can support VCM SDRAM, including Intel's 815E and VIA's 694X.
- After the introduction of Intel: PC-100, due to the development of technology, the 800MB / s bandwidth of PC-100 memory can no longer meet the demand, while the bandwidth of PC-133 does not increase much (1064MB / s), and it cannot meet future developments. demand. In order to achieve the purpose of monopolizing the market, Intel and Rambus jointly promoted Rambus DRAM (DirectRambus DRAM) in the PC market.
- Rambus uses a 400MHz 16bit bus. In one clock cycle, data can be transmitted at the same time on the rising and falling edges. In this way, its actual speed is 400MHz × 2 = 800MHz, and the theoretical bandwidth is (16bit × 2 × 400MHz / 8) 1.6GB / s, equivalent to twice the PC-100. In addition, Rambus can also store 9-bit bytes. The extra bit is reserved and may be used as an ECC (ErroI · Checking and Correction) check bit in the future. The Rambus clock can be up to 400MHz, and only 30 copper wires are used to connect the memory controller and RIMM (Rambus In-line Memory Modules, Rambus embedded memory modules). Reducing the length and number of copper wires can reduce data transmission Electromagnetic interference, which can quickly increase the operating frequency of the memory. However, at high frequencies, the heat it emits will definitely increase, so the first Rambus memory even needs its own cooling fan.
- 3.DDR3 era
- Compared with DDR2, DDR3 has a lower operating voltage, which has dropped from 1.8V to 1.5V for DDR2, which has better performance and saves power; 4bit pre-reading of DDR2 is upgraded to 8-bit prereading. DDR3 currently can reach a maximum speed of 2000Mhz. Although the fastest DDR2 memory speed has been increased to 800Mhz / 1066Mhz, the DDR3 memory module will still jump from 1066Mhz.
- First, the new design of DDR3 based on DDR2:
- 1. 8bit prefetch design, and DDR2 is 4bit prefetch, so the frequency of DRAM core is only 1/8 of the interface frequency, and the core operating frequency of DDR3-800 is only 100MHz.
- 2. A point-to-point topology is used to reduce the burden on the address / command and control bus.
- 3 Using the production process below 100nm, the operating voltage is reduced from 1.8V to 1.5V, and asynchronous reset (Reset) and ZQ calibration functions are added.