Differences among DRAM, SDRAM, and SRAM

Guide: Differences among DRAM, SDRAM, and SRAM: SRAM is short for static RAM without refreshing, whose speed is very fast. The cache inside the CPU belongs to static RAM. However, the disadvantage is that a memory unit requires a large number of transistors, so it is expensive in spite of small capacity.

Differences among DRAM, SDRAM, and SRAM

Brief introduction of features:

SRAM: SRAM is short for static RAM without refreshing, whose speed is very fast. The cache inside the CPU belongs to static RAM. However, the disadvantage is that a memory unit requires a large number of transistors, so it is expensive in spite of small capacity.

DRAM: Dynamic RAM requires refreshing, with large capacity.

SDRAM: Synchronous dynamic RAM need to refresh, with fast speed and large capacity.

DDR SDRAM: Dual-channel synchronous dynamic RAM, requires refreshing, with fast speed and large capacity.

Specific explanation one:

What is DRAM?

DRAM is called Dynamic RAM. DRAM is used for general data access. What we often say how much memory is mainly refers to the capacity of DRAM.

What is SRAM?

SRAM is short for Static RAM, mainly used to make cache.

What is SDRAM?

SDRAM is known as Synchronous DRAM, faster than the general DRAM and EDO RAM, which has gradually become the standard memory configuration of the PC.

What is Cache?

The English meaning of Cache is storage, which is generally manufactured by using SRAM. Cache exchanges data with CPU, faster than DRAM, so it is called high-speed cache memory, abbreviated high-speed cache. Because the information processing speed of CPU often exceeds that ofother parts, the use of general DRAM as information memory often puts the CPU in a waiting state, resulting in the waste of resources. Cache was born to solve this problem. After the operating system starts, CPU temporarily stores some of the system information that is often stored in DRAM in the cache. Later when CPU needs to use this information, first go to the cache to find. If found, read it directly from the cache, contributing to save a lot of time by taking advantage of the high-speed performance. Most CPUs integrate a certain amount of cache into themselves, commonly referred to as primary cache or built-in cache. The speed of exchanging information between this part of memory and CPU is the fastest, but its capacity is small. Most motherboards are also integrated with cache, commonly referred to as secondary cache or external cache, larger than the built-in cache in capacity, generally up to 256K. Now some motherboards have already used high-speed cache with capacity of 512K～2M. The latest Pentium II CPU is integrated with a primary cache and l secondary cache, and at that time the cache on the motherboard can only be called third cache.

What is flash memory?

Flash memory is currently used in most BIOS on the motherboard, which can upgrade BIOS directly by adjusting the voltage on the motherboard.

Explain why DRAM need to refresh, SRAM not:

This depends on the type of RAM design. DRAM uses a T and a RC circuit, resulting in capacitance destruction leakage and slow discharge. So you need to refresh frequently to keep the data.

Specific explanation two: 

DRAM, short for dynamic random access memory, requires constant refresh to save data. Its row and column addresses multiplex. Many of DRAM have page mode.

SRAM, short for static random access memory, does not need to refresh when power is on, without data lost. Moreover, its row and column addresses generally doesn’t multiplex.

SDRAM, short for synchronous DRAM, means reading and writing data requires a clock to synchronize.

Due to the different implementation process, DRAM and SDRAM have larger capacity than SRAM, while the speed of reading and writing is not as good as SRAM. But now, SDRAM is very fast, and its clock seems to have 150 megabytes with reading and writing cycle less than 10ns. Although SDRAM works at a high frequency, the actual throughput rate should be reduced. Taking PC133 as an example, its clock cycle is 7.5ns. When CAS latency = 2, it needs 12 cycles to complete 8 burst reads and 10 cycles to complete 8 burst writes. However, if the Band is accessed alternately, SDRAM can complete a read-write operation (of course except for refreshing). In fact, the current mainstream high-speed memory is SSRAM (synchronous SRAM) and SDRAM (synchronous DRAM). At present, the maximum capacity of SSRAM that can be easily bought is 8Mb per chip, with the maximum working speed of 166MHz. The maximum capacity of SDRAM that can be easily bought is 128Mb per chip, with the maximum working speed of 133MHz.

SRAM, an abbreviation for Static Random Access Memory, is a type of semiconductor memory. "Static" means that data stored in SRAM is not lost as long as power is not turned off. This is different from dynamic RAM (DRAM), which requires periodic refresh operations. Of course, we should not confuse SRAM with ROM and flash memory, because SRAM is a volatile memory that can maintain data only if the power supply remains continuously supplied. "Random access" means that the contents of the memory can be accessed in any order, regardless of the location of the previous access. 

Each bit of SRAM is stored in four transistors, which make up two cross-coupled inverters. This storage unit has two stable states, usually expressed as 0 and 1. Two access transistors are required to control access to storage units during read or write operations. Therefore, a storage bit typically requires six MOSFET. The symmetrical circuit structure allows SRAM to access faster than DRAM. Another reason why SRAM is faster than DRAM is that SRAM can receive all address bits at once, while DRAM uses the structure of row and column address multiplexing.

SRAM should not be confused with SDRAM. SDRAM represents synchronous DRAM, which is completely different from SRAM. SRAM should also not be confused with PSRAM, which is a kind of DRAM disguised as SRAM.

From the type of transistor, SRAM can be divided into bipolarity and CMOS. Functionally, SRAM can be divided into asynchronous SRAM and synchronous SRAM. The access of asynchronous SRAM is independent of the clock, when the input and output of data are controlled by the change of address. All access to synchronous SRAM is initiated at the rising/falling edge of the clock. Addresses, data inputs, and other control signals are all related to the clock signals.

DRAM, short for dynamic random access memory, requires constant refresh to save data. Its row and column addresses multiplex. Many of DRAM have page mode.

SRAM, short for static random access memory, does not need to refresh when power is on, without data lost. Moreover, its row and column addresses generally doesn’t multiplex.

SDRAM, short for synchronous DRAM, means reading and writing data requires a clock to synchronize.

Different storage unit structure results in different capacity. A DRAM storage unit requires approximately one transistor and one capacitor (excluding line readout amplifiers, etc.), while a SRAM storage unit needs about six transistors. Due to the different implementation process, DRAM and SDRAM have larger capacity than SRAM, while the speed of reading and writing is not as good as SRAM.

One is static, the other is dynamic. The static is to use bistable triggers to save information, while the dynamic is to use electrons refreshing from time to time.

RAM can be divided into SRAM and DRAM. "Memory" we often say refers to DRAM while SRAM is rarely touched.

SRAM is actually a very important memory, which has a wide range of uses. SRAM is very fast and maintains data integrity when it is read and refreshed quickly. Due to the form of bistable circuit inside SRAM to store data, the circuit structure of SRAM is very complex. The cost of making SRAM with the same capacity is much higher than that of DRAM. Because of this, its development has been limited. As a result, SRAM is currently used only for primarily cache within the CPU and the built-in secondary cache. Only a small number of network servers and routers can use SRAM.