摘要:
A Dynamic Random Access Memory (DRAM) performs read, write, and refresh operations. The DRAM includes a plurality of sub-arrays, each having a plurality of memory cells, each of which is coupled with a complementary bit line pair and a word line. The DRAM further includes a word line enable device for asserting a selected one of the word lines and a column select device for asserting a selected one of the bit line pairs. A timing circuit is provided for controlling the word line enable device, the column select device, and the read, write, and refresh operations in response to a word line timing pulse. The read, write, and refresh operation are performed in the same amount of time.
摘要:
An image memory, image memory system, memory controller and semiconductor integrated circuit that are capable of efficiently accessing a rectangular area of two-dimensionally arrayed data are provided. The semiconductor integrated circuit in which circuits are integrated on a single semiconductor substrate, the semiconductor integrated circuit comprising: a memory array (224) in which memory cells connected to word lines and bit lines and storing data are arranged in the form of a matrix, and which is sectioned into bit groups for each of Nb number of memory cells; and a controller that selects a bit group as a starting point, on the basis of the address signal and first combination information, the controller outputting, to Nb x N number of output terminals (DQ) in parallel, on the basis of second combination information, data items that are stored in the respective memory cells contained in N number of bit groups including the bit group as the starting point.
摘要:
Typically, a bulk of the memory space utilized by an SOC (103) is located in cheaper off-chip memory devices such as Synchronous Dynamic Random Access Memory (SDRAM) memories (104). These memories provide a large capacity for data storage, at a relatively low cost. It is common for SOC devices to communicate with each other through these off-chip memory devices. Because of the large amount of data being processed on state of the art SOCs, the data bandwidth to and from the SDRAM memories is a critical resource, which if improperly managed results in bottlenecks. Thus, a novel address mapping scheme, which has improved efficiency for two-dimensional memory transactions is proposed for mapping on-chip memory transactions to off-chip memory transactions. This novel mapping scheme aims to decrease these bottlenecks, by segmenting the data sequence into portions being smaller than the size of a row of a block of the SDRAM memories.
摘要:
An image memory, image memory system, and memory controller that are capable of efficiently accessing a rectangular area of two-dimensionally arrayed data are provided. The memory device has: a memory cell array that has a plurality of memory unit areas, each of which is selected by addresses; a plurality of input/output terminals; and an input/output unit provided between the memory cell array and the plurality of input/output terminals. Each of the memory unit areas stores therein data of a plurality of bytes or bits corresponding to the plurality of input/output terminals respectively, and the memory cell array and the input/output unit access a plurality of bytes or bits stored in a first memory unit area corresponding to the input address and in a second memory unit area adjacent to the first memory unit on the basis of the input address and combination information of the bytes or bits in response to a first operation code, and then, from the plurality of bytes or bits within the accessed first and second memory unit areas, associate a combination of the plurality of bytes or bits based on the combination information, with the plurality of input/output terminals.
摘要:
Described is a memory system in which the memory core organization changes with device width. The number of physical memory banks accessed reduces with device width, resulting in reduced power usage for relatively narrow memory configurations. Increasing the number of logic memory banks for narrow memory widths reduces the likelihood of bank conflicts, and consequently improves speed performance.
摘要:
A method and system for maximizing DRAM memory bandwidth is provided. The system includes a plurality of buffers to store a plurality of data units, a selector coupled to the buffers to select the buffer to which a data unit is to be stored, and logic coupled to the buffers to schedule an access of one of a corresponding number of memory banks based on the buffer in which the data unit is stored. The system receives a data unit, computes an index based on at least a portion of the data unit, selects a buffer in which to store the data unit based on the index, stores the data unit in the selected buffer, schedules a memory bank access based on the index, reads the data unit from the selected buffer, and accesses the memory bank.
摘要:
A semiconductor memory is provided with a plurality of first memory blocks and a second memory block for reproducing data of the first memory blocks. When a read command and a refresh command conflict with each other, a read control circuit accesses the first memory block according to the refresh command and reproduces read data by using the second memory block. When a write command (WRA) and the refresh command (REFRQ) conflict with each other, a write control circuit operates the memory block according to an order of command reception. Therefore, it is possible to perform refresh operation without being recognized by users. Namely, a user-friendly semiconductor memory can be provided.
摘要:
A high-speed clock-synchronous memory device is provided with a sense amplifier (S/A) shared by and between cell arrays, and a cell array controller unit (CNTRLi), wherein input and output of data/command synchronous with the clock, access command supplies all address data bits (row and column) simultaneously. By acknowledging a change in bits observed between two successive commands, regarding some of address bits configuring access address, the device judges whether the current access is made within the same cell array as the preceding access (S), between the neighboring cell arrays (N), or between remote cell arrays (F). According to the judgement, suitable command cycle is applied. At this time, the command cycle satisfies relationship: S ≧ N ≧ F.