摘要:
A memory device includes multiple data I/O lanes and corresponding lane or column decoders. Instead of providing the same address to each column decoder, decoder logic calculates potentially different column addresses depending on the needs of the device utilizing the memory. For example, the column addresses might be based on a received CAS address and an accompanying offset. This allows data access at alignments that do not necessarily correspond to CAS alignments. The technique is utilized in conjunction with graphics systems in which tiling is used. In systems such as this, memory offsets are specified in terms of pixel columns and rows. The technique is also used in conjunction with a router such as a TCP/IP router, in which individual packets are aligned at CAS boundaries. In this situation, the decoder logic is alternatively configurable to allow access of either an information packet or a plurality of packet headers during a single memory access cycle.
摘要:
Methods and arrangements are provided for use in memory devices, which allow column address strobe (CAS) timing to adjust to, and/or be adjusted by a controller to, have both minimal unloaded latency and optimal pipelined latency. A delay CAS (DC) period is only applied until a row-to-column delay (tRCD) has been satisfied. Once the tRCD has been satisfied, then the DC period is not enforced for subsequent CAS operations within the memory core associated with a page hit. When a subsequent read command is received at the input/output pins of the memory device and a corresponding RAS operation is performed in the memory core, then the tRCD will again need to be satisfied and a DC period will again be enforced. Consequently the methods and arrangements allow the CAS delay to be dynamically and selectively adjusted to best support the workload. This results in better performance and increased bandwidth.
摘要:
An apparatus for controlling a dynamic random access memory (DRAM), the apparatus comprising an interface to transmit, over a first plurality of wires, to the DRAM a first code to indicate that first data is to be written to the DRAM and a column address to indicate a column location of a memory core in the DRAM where the first data is to be written. The interface is further to transmit a second code to indicate whether mask information for the first data will be sent to the DRAM. If the second code indicates that mask information will be sent, a portion of the column address and a portion of the mask information are sent after the second code is sent. The interface is further to transmit to the DRAM, over a second plurality of wires separate from the first plurality of wires, the first data.
摘要:
A DRAM controller component generates a timing signal and transmits, to a DRAM, write data that requires a first time interval to propagate from the DRAM controller component to the DRAM and to be sampled by the DRAM on one or more edges of the timing signal, a clock signal that requires a second time interval to propagate from the DRAM controller component to the DRAM, and a write command, associated with the write data, to be sampled by the DRAM on one or more edges of the clock signal. The DRAM controller component includes series-coupled delay elements to generate respective incrementally delayed signals, and a multiplexer to select one of the delayed signals to time the transmission of the write data, such that transmission of the write data is delayed based on a difference between the first time interval and the second time interval.
摘要:
A plurality of control signals are asserted within an asynchronous integrated circuit memory device in response to each transition of a memory access initiation signal to effect pipelined memory access operations.
摘要:
An apparatus for controlling a dynamic random access memory (DRAM), the apparatus comprising an interface to transmit to the DRAM a first code to indicate that first data is to be written to the DRAM. The first code is to be sampled by the DRAM and held by the DRAM for a first period of time before it is issued inside the DRAM. The interface is further to transmit the first data that is to be sampled by the DRAM after a second period of time has elapsed from when the first code is sampled by the DRAM. The interface is further to transmit a second code, different from the first code, to indicate that second data is to be read from the DRAM. The second code is to be sampled by the DRAM on one or more edges of the external clock signal.
摘要:
A memory controller is disclosed. In one particular exemplary embodiment, the memory controller may comprise a first transmitter to output first and second write commands synchronously with respect to a clock signal, a second transmitter to output first data using a first timing offset such that the first data arrives at a first memory device in accordance with a predetermined timing relationship with respect to a first transition of the clock signal, and a third transmitter to output second data suing a second timing offset such that the second data arrives at a second memory device in accordance with a predetermined timing relationship with respect to a second transition of the clock signal.
摘要:
This disclosure provides a method for adjusting system timing in a reconfigurable memory system. In a Dynamic Point-to-Point (“DPP”) system, for example, manufacturer-supplied system timing parameters such as access latency and maximum clock speed typically reflect a worst-case configuration scenario. By in-situ detecting actual configuration (e.g., whether expansion boards have been inserted), and correspondingly configuring the system to operate in a mode geared to the specific configuration, worst-case or near worst-case scenarios may be ruled out and system timing parameters may be redefined for faster-than-conventionally-rated performance; this is especially the case in a DPP system where signal pathways typically become more direct as additional modules are added. Contrary to convention wisdom therefore, which might dictate that component expansion should slow down timing, clock speed can actually be increased in such a system, if supported by the configuration, for better performance.
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:
In a method of controlling a memory device, the following is conveyed over a first set of interconnect resources: a first command that specifies activation of a row of memory cells; a second command that specifies a write operation, wherein write data is written to the row; a bit that specifies whether precharging occurs after the write data is written; and a code that specifies whether data mask information will be issued for the write operation. If the code specifies that the information will be issued, then the information, which specifies whether to selectively write portions of the write data, is conveyed over the first set of interconnect resources after conveying the code. The write data to be written in connection with the write operation is conveyed over a second set of interconnect resources that is separate from the first set of interconnect resources.