Abstract:
The present invention relates to a system and method for processing the read and write operations in a memory architecture. The system processing the read and write operations includes at least one local memory block and a synchronously controlled global controller coupled to the local memory block and adapted to extend the high portion of a clock pulse. The method for processing the read and write operations includes skewing a clock pulse using at least one word line interfacing with the global controller.
Abstract:
In one embodiment, a decoder for decoding an address having a plurality of bits ranging from a first address bit a1 to a last address bit aN, each address bit being either true or false is provided that includes: a pre-charge circuit adapted to pre-charge a dynamic NOR node and a dynamic OR node and then allow the pre-charged dynamic NOR node and pre-charged dynamic OR node to float; a plurality of switches coupled between the dynamic NOR node and ground, each switch corresponding uniquely to the address bits such that the switches range from a first switch corresponding to a1 to an nth switch corresponding to aN, wherein any switch corresponding to a true address bit is configured to turn on only if its corresponding address bit is false, and wherein any switch corresponding to a false address bit is configured to turn on only if its corresponding address bit is true; a (n+1)th switch coupling the dynamic OR node to ground, the (n+1)th switch being controlled such that it turns on if the dynamic OR node is charged, whereby the pre-charged dynamic OR node discharges if the dynamic OR node remains charged; an odd plurality of inverters coupled in series with the dynamic OR node; and a word line driven by the odd plurality of inverters.
Abstract:
In one embodiment, a method of leakage control for a memory having an array of memory cells arranged into a plurality of sub-arrays is provided wherein each sub-array has a sleep mode of operation controlled by a sleep signal in which stored data is lost, and wherein each sub-array asserts a local clock if the sub-array is addressed. The method includes the act of asserting a sleep signal while addressing a given one of the sub-arrays such that only the given one of the sub-arrays is placed into the sleep mode.
Abstract:
In one embodiment, a multi-port SRAM is provided that comprises: a single input port and output port 6-T SRAM; and a multi-port control block circuit that includes: a plurality of input registers corresponding to a plurality of input ports to register corresponding input signals; an input multiplexer to select from the input registers to provide a selected input signal to the 6-T SRAM's single input port; a plurality of output registers corresponding to a plurality of output ports to register corresponding output signals; and an output de-multiplexer to select from the output registers to provide an output signal from the 6-T SRAM's single output port to the selected output register.
Abstract:
A one-time-programmable memory cell uses two complementary antifuses that are programmed in a complementary fashion such that only one of the two complementary antifuses is stressed by a programming voltage. The programming voltage stress one a particular one of the complementary antifuses indicates a logical state of the memory cell. For example, a logical high state may correspond to a first one of the complementary antifuses being stressed whereas a logical low state may correspond to the stressing of the remaining one of the complementary antifuses.
Abstract:
Aspects of the invention may include testing and debugging an embedded device under test. Testing and debugging and embedded device under test may include the step of loading an instruction into a parameterized shift register of each one of a plurality of BIST modules coupled to an individual one of a plurality of embedded memory modules comprising the embedded device under test. An identity of each of the instruction loaded into the parameterized shift register of each one of the plurality of BIST modules may subsequently be determined. A separate test signal may be generated from each one of the plurality of BIST modules corresponding to the determined identity of the instruction loaded in each one of the plurality of BIST modules, each one of the generated test signals causing control and execution of the testing and debugging of a corresponding one of each of the plurality of embedded memory modules comprising the embedded device under test.
Abstract:
In one embodiment, a hybrid DRAM is provided that includes: a sense amplifier including a differential amplifier and regenerative latch, wherein the differential amplifier and regenerative latch are constructed using core transistors; and a plurality of memory cells coupled to the sense amplifier through a pair of bit lines, wherein the memory cells are constructed using I/O transistors.
Abstract:
In one embodiment, a DRAM is provided that includes a plurality of memory cells, each memory cell including an access transistor and a storage capacitor, wherein the storage capacitor includes a first node coupled to the access transistor and a second node isolated from the first node, the second node comprising signal-bearing metal conductors.
Abstract:
In one embodiment, a memory includes a row and/or column redundancy architecture that uses binary cells to indicate whether a given row or column of memory cells is faulty. The binary cell is adapted to store a “repair true” signal in response to a conventional access to the corresponding row or column and also the assertion of a set signal.
Abstract:
In one embodiment, NMOS transistors have their source coupled to a common source node such that the NMOS transistors conduct a leakage current if the common source node is grounded. To reduce this leakage current, the common source node is raised in potential. Similarly, PMOS transistors have their source coupled to a common source node such that the PMOS transistors conduct a leakage current if the common source node is charged to a power supply voltage VDD. To reduce this leakage current, the common source node is lowered in potential.