The fundamental difference between SRAM and DRAM

In the storage system,SRAM,DRAM,Flash and other storage technologies are not absolutely superior or inferior.The core is to choose between speed,density and cost and match the design goals of different system scenarios.
 

SRAM relies on bistable latch structure composed of cross-coupled inverters to register data.The mainstream is 6T cell architecture,and there are also 4T and other derivative schemes.As long as the power is continuously supplied,the latch can stably maintain the storage state,without periodic refresh operation,with lower access delay and fast response speed.It is the first choice for CPU cache,on-chip storage and other scenes that require strict real-time and low delay.
 

DRAM adopts classic 1T1C cells,and relies on capacitors to store charge characterization data.The capacitor naturally has charge leakage,so it must be refreshed regularly,otherwise the data will be lost;However,the cell structure is very simple,occupies less wafer area,and has obvious advantages in storage density and single-bit cost,so it is suitable for building large-capacity memory.
 

The hierarchical storage architecture of modern computers is precisely to reconcile the contradiction between them:using SRAM nearby to reduce access delay,relying on DRAM to carry large-capacity data,combining the system running speed and storage cost,and achieving the best performance of the whole machine.