Showing papers on "Registered memory published in 2002"

Selective operation of a multi-state non-volatile memory system in a binary mode

Abstract: A flash non-volatile memory system that normally operates its memory cells in multiple storage states is provided with the ability to operate some selected or all of its memory cell blocks in two states instead. The two states are selected to be the furthest separated of the multiple states, thereby providing an increased margin during two state operation. This allows faster programming and a longer operational life of the memory cells being operated in two states when it is more desirable to have these advantages than the increased density of data storage that multi-state operation provides.

Dual-port buffer-to-memory interface

Abstract: Methods and apparatus for a memory system using a new memory module architecture are disclosed. In one embodiment, the memory module has two ranks of memory devices, each rank connected to a corresponding one of two 64-bit-wide data registers. The data registers connect to two 64-bit-wide ports of a 120:64 multiplexer/demultiplexer, and a 64-bit-wide data buffer connects to the opposite port of the multiplexer/demultiplexer. A controller synchronizes the operation of the data registers, the multiplexer/demultiplexer, and the data buffer. In an operating environment, the data buffer connects to a memory bus. When a data access is performed, both ranks exchange data signaling with their corresponding data registers during a single data access. At the buffer, the memory bus data transfer occurs in two consecutive clock cycles, one cycle for each rank. This allows the memory bus transfer rate to double for the same memory bus width and memory device speed.

An optimal memory allocation scheme for scratch-pad-based embedded systems

Abstract: This article presents a technique for the efficient compiler management of software-exposed heterogeneous memory. In many lower-end embedded chips, often used in microcontrollers and DSP processors, heterogeneous memory units such as scratch-pad SRAM, internal DRAM, external DRAM, and ROM are visible directly to the software, without automatic management by a hardware caching mechanism. Instead, the memory units are mapped to different portions of the address space. Caches are avoided due to their cost and power consumption, and because they make it difficult to guarantee real-time performance. For this important class of embedded chips, the allocation of data to different memory units to maximize performance is the responsibility of the software.Current practice typically leaves it to the programmer to partition the data among different memory units. We present a compiler strategy that automatically partitions the data among the memory units. We show that this strategy is optimal, relative to the profile run, among all static partitions for global and stack data. For the first time, our allocation scheme for stacks distributes the stack among multiple memory units. For global and stack data, the scheme is provably equal to or better than any other compiler scheme or set of programmer annotations. Results from our benchmarks show a 44.2p reduction in runtime from using our distributed stack strategy vs. using a unified stack, and a further 11.8p reduction in runtime from using a linear optimization strategy for allocation vs. a simpler greedy strategy; both in the case of the SRAM size being 20p of the total data size. For some programs, less than 5p of data in SRAM achieves a similar speedup.

Non-volatile semiconductor memory device adapted to store a multi-valued data in a single memory cell

Abstract: A non-volatile semiconductor memory device includes an electrically data rewritable non-volatile semiconductor memory cell and a write circuit for writing data in the memory cell, the write circuit writing a data in the memory cells by supplying a write voltage Vpgm and a write control voltage VBL to the memory cell, continuing the writing of the data in the memory cell by changing the value of the write control voltage VBL in response to an advent of a first write state of the memory cell and inhibiting any operation of writing a data to the memory cell by further changing the value of the write control voltage VBL to Vdd in response to an advent of a second write state of the memory cell.

Mondrian memory protection

Abstract: Mondrian memory protection (MMP) is a fine-grained protection scheme that allows multiple protection domains to flexibly share memory and export protected services. In contrast to earlier page-based systems, MMP allows arbitrary permissions control at the granularity of individual words. We use a compressed permissions table to reduce space overheads and employ two levels of permissions caching to reduce run-time overheads. The protection tables in our implementation add less than 9% overhead to the memory space used by the application. Accessing the protection tables adds than 8% additional memory references to the accesses made by the application. Although it can be layered on top of demand-paged virtual memory, MMP is also well-suited to embedded systems with a single physical address space. We extend MMP to support segment translation which allows a memory segment to appear at another location in the address space. We use this translation to implement zero-copy networking underneath the standard read system call interface, where packet payload fragments are connected together by the translation system to avoid data copying. This saves 52% of the memory references used by a traditional copying network stack.

Method for fabricating programmable memory array structures incorporating series-connected transistor strings

Abstract: A three-dimensional flash memory array incorporates thin film transistors having a charge storage dielectric arranged in series-connected NAND strings to achieve a 4F2 memory cell layout. The memory array may be programmed and erased using only tunneling currents, and no leakage paths are formed through non-selected memory cells. Each NAND string includes two block select devices for respectively coupling one end of the NAND string to a global bit line, and the other end to a shared bias node. Pairs of NAND strings within a block share the same global bit line. The memory cells are preferably depletion mode SONOS devices, as are the block select devices. The memory cells may be programmed to a near depletion threshold voltage, and the block select devices are maintained in a programmed state having a near depletion mode threshold voltage. NAND strings on more than one layer may be connected to global bit lines on a single layer. By interleaving the NAND strings on each memory level and using two shared bias nodes per block, very little additional overhead is required for the switch devices at each end of the NAND strings. The NAND strings on different memory levels are preferably connected together by way of vertical stacked vias, each preferably connecting to more than one memory level. Each memory level may be produced with less than three masks per level.

Non-volatile memory control

Abstract: According to an embodiment of the present invention, there is provided a method and apparatus for use in a memory system having a non-volatile memory and a controller for limiting the number of non-volatile memory arrays from a plurality of available arrays accessed at one time, wherein the method comprises implementing a pipelining sequence for transferring data to and from the non-volatile memory arrays and limiting the number of active arrays operating at one time, the arrangement being such that the controller waits for the at least one of the arrays to complete before initiating the transfer to and from a further array.

Safe memory reclamation for dynamic lock-free objects using atomic reads and writes

Abstract: A major obstacle to the wide use of lock-free data structures, despite their many performance and reliability advantages, is the absence of a practical lock-free method for reclaiming the memory of dynamic nodes removed from dynamic lock-free objects for arbitrary reuse.The only prior lock-free memory reclamation method depends on the DCAS atomic primitive, which is not supported on any current processor architecture. Other memory management methods are blocking, require special operating system support, or do not allow arbitrary memory reuse.This paper presents the first lock-free memory management method for dynamic lock-free objects that allows arbitrary memory reuse, and does not require special operating system or hardware support. It guarantees an upper bound on the number of removed nodes not yet freed at any time, regardless of thread failures and delays. Furthermore, it is wait-free, it is only logarithmically contention-sensitive, and it uses only atomic reads and writes for its operations. In addition, it can be used to prevent the ABA problem for pointers to dynamic nodes in most algorithms, without requiring extra space per pointer or per node.

Method and apparatus for using smart memories in computing

Abstract: A smart memory computing system that uses smart memory for massive data storage as well as for massive parallel execution is disclosed. The data stored in the smart memory can be accessed just like the conventional main memory, but the smart memory also has many execution units to process data in situ. The smart memory computing system offers improved performance and reduced costs for those programs having massive data-level parallelism. This smart memory computing system is able to take advantage of data-level parallelism to improve execution speed by, for example, use of inventive aspects such as algorithm mapping, compiler techniques, architecture features, and specialized instruction sets.

Enhanced memory module architecture

Abstract: A memory module architecture that supports Flash and static memory devices in addition to dynamic memory devices. The module architecture of the present invention preferably redefines standard application of chip select signals on existing module architectures to provide requisite signaling to support Flash and static RAM devices. Use of serial presence detect signaling features of standard memory modules is also modified to provide desired identity and parameters of such an enhanced module. A further aspect of the present invention provides for a support structure to immobilize the loose edge of the memory module opposite the electrical edge connector of the module to further enhance the module's resistance to vibration and mechanical shock by immobilizing the module with respect to rotation about the key in the socket connector.

Memory module having buffer for isolating stacked memory devices

Abstract: The present invention utilizes a buffer to isolate a stack of memory devices, thereby taking advantage of the increased memory density available from stacked memory devices while reducing capacitive loading. A memory module in accordance with the present invention may include a stack of memory devices and a buffer coupled to the first and second memory devices and arranged to capacitively isolate the first and second memory devices from a bus. In a memory system in accordance with the present invention, multiple buffered stacks of memory devices are preferably coupled in a point-to-point arrangement, thereby further reducing capacitive loading.

Method of writing data to non-volatile memory

Abstract: According to a first aspect of the invention, there is provided a controller connected to a non-volatile memory and including a volatile memory, wherein the controller maintains lists in volatile memory of blocks in the non-volatile memory allocated for storage of logical sector data and of blocks recently erased in the non-volatile memory.

Fail-over of multiple memory blocks in multiple memory modules in computer system

Abstract: A computer system has a memory controller for controlling accesses to multiple memory modules, each having multiple memory blocks, and a fail-over circuit for failing-over individual memory blocks from multiple memory modules. The digital information stored in an individual memory block that has experienced memory errors in excess of a permissible threshold is copied to an auxiliary memory location. The memory accesses directed to the failed-over memory block are intercepted and redirected to the auxiliary memory location. Tags are stored to identify failed-over memory modules and corresponding auxiliary memory modules, so a tag look-up for an accessed memory address can generate a hit signal when the memory access is to a failed-over memory module and cause the auxiliary memory module to respond to the memory access.

Process for automatic dynamic reloading of data flow processors (dfps) and units with two-or-three-dimensional programmable cell architectures (fpgas, dpgas, and the like)

Abstract: In a data-processing method, first result data may be obtained using a plurality of configurable coarse-granular elements, the first result data may be written into a memory that includes spatially separate first and second memory areas and that is connected via a bus to the plurality of configurable coarse-granular elements, the first result data may be subsequently read out from the memory, and the first result data may be subsequently processed using the plurality of configurable coarse-granular elements. In a first configuration, the first memory area may be configured as a write memory, and the second memory area may be configured as a read memory. Subsequent to writing to and reading from the memory in accordance with the first configuration, the first memory area may be configured as a read memory, and the second memory area may be configured as a write memory.

Memory card and memory controller

Abstract: A memory card has a plurality of non-volatile memories and a main controller for controlling the operation of the non-volatile memories. The main controller performs an access control to the non-volatile memories in response to an external access instruction, and an alternate control for alternating an access error-related storage area of the non-volatile memory with other storage area. In the access control, the speeding up of the data transfer between flash memories is achieved by causing the plurality of non-volatile memories to parallel access operate. In the alternation control, the storage areas is made alternative for each non-volatile memory in which an access error occurs.

Laser programmable electrically readable phase-change memory method and device

Abstract: Roughly described, a phase-change memory such as a chalcogenide-based memory is programmed optically and read electrically. No complex electrical circuits are required for programming the cells. On the other hand, this memory can be read by electrical circuitry directly. The read out speed is much faster than for optical disks, and integrated circuit chips made this way are more compatible with other electrical circuits than are optical disks. Thus memories according to the invention can have simple, low power-consuming, electrical circuits, and do not require slow and power-hungry disk drives for reading. The invention therefore provides a unique low power, fast read/write memory with simple electrical circuits.

Memory system and memory card

Abstract: A memory system comprises nonvolatile memory chips (CHP1, CHP2) having memory banks (BNK1, BNK2) capable of performing memory operations independently and a memory controller (5) capable of accessing/controlling separately the nonvolatile memory chips. The memory controller can selectively instruct the memory banks of the nonvolatile memory chips to perform a simultaneous or interleave write operation. Therefore, the simultaneous write operations each requiring a write time much longer than the write set-up time can be completely parallel carried out, and the interleave write operations following the write set-up can be carried out parallel and overlapped with a write operation of another memory bank. As a result, the number of nonvolatile memory chips constituting a memory system capable of performing a high-speed write operation can be relatively small.

On a new way to read data from memory

Abstract: This paper explains a new family of techniques to extract data from semiconductor memory, without using the read-out circuitry provided for the purpose. What these techniques have in common is the use of semi-invasive probing methods to induce measurable changes in the analogue characteristics of the memory cells of interest. The basic idea is that when a memory cell, or read-out amplifier, is scanned appropriately with a laser, the resulting increase in leakage current depends on its state; the same happens when we induce an eddy current in a cell. These perturbations can be carried out at a level that does not modify the stored value, but still enables it to be read out. Our techniques build on it number of recent advances in semi-invasive attack techniques, low temperature data remanence, electromagnetic analysis and eddy current induction. They can be used against a wide range of memory structures, from registers through RAM to FLASH. We have demonstrated their practicality by reading out DES keys stored in RAM without using the normal read-out circuits. This suggests that vendors of products such as smartcards and secure microcontrollers should review their memory encryption, access control and other storage security issues with care.

Disk drive employing a non-volatile serial semiconductor memory for storing a control program for a microprocessor

Abstract: A disk drive is disclosed comprising a microprocessor, a non-volatile serial semiconductor memory for storing code segments of a control program, a first semiconductor memory for storing code segments of a loader program, and a second semiconductor memory for receiving the code segments of the control program. When the disk drive is powered on, the microprocessor executes the loader program from the first semiconductor memory to load the control program from the non-volatile serial semiconductor memory into the second semiconductor memory. The microprocessor then executes the control program from the second semiconductor memory.

Method and apparatus for memory access scheduling to reduce memory access latency

Abstract: A device is presented including a memory controller. The memory controller is connected to a read request queue. A command queue is coupled to the memory controller. A memory page table is connected to the memory controller. The memory page table has many page table entries. A memory page history table is connected to the memory controller. The memory history table has many page history table entries. A pre-calculated lookup table is connected to the memory controller. The memory controller includes a memory scheduling process to reduce memory access latency.

Appliance, including a flash memory, that is robust under power failure

Abstract: An appliance that includes a host device and a memory unit with a primary memory, and a method of operating the appliance. According to one aspect of the appliance, the primary memory is nonvolatile and the memory unit also includes a volatile memory a power sensor and a controller. When the power sensor detects interruption of power to the memory unit, the controller copies data selectively from the volatile memory to the primary memory. Power for this copying is provided by a secondary power source such as a battery or a capacitor. According to another aspect of the appliance, the appliance includes primary and secondary power sources, and the memory unit also includes a charge pump whose functions include both boosting power from the primary source for the primary memory and charging the secondary source.

Flash memory with multiple status reading capability

Abstract: A synchronous flash memory includes an array of non-volatile memory cells. The memory array is arranged in rows and columns, and can be further arranged in addressable blocks. Data communication connections are used for bi-directional data communication with an external device(s), such as a processor or other memory controller. A write latch is coupled between the data buffer and the memory array to latch data provided on the data communication connections. The memory can write data to one location, such as a memory array block, while data is read from a second location, such as a second memory array block. The memory automatically provides status data when a read command is received for a memory array location that is currently subject to a write operation. The automatic status output allows multiple processors to access the memory device without substantial bus master overhead. The memory can also output status data in response to a status read command.

Programmable memory array structure incorporating series-connected transistor strings and methods for fabrication and operation of same

Abstract: A three-dimensional flash memory array incorporates thin film transistors having a charge storage dielectric arranged in series-connected NAND strings to achieve a 4F2 memory cell layout. The memory array may be programmed and erased using only tunneling currents, and no leakage paths are formed through non-selected memory cells. Each NAND string includes two block select devices for respectively coupling one end of the NAND string to a global bit line, and the other end to a shared bias node. Pairs of NAND strings within a block share the same global bit line. The memory cells are preferably depletion mode SONOS devices, as are the block select devices. The memory cells may be programmed to a near depletion threshold voltage, and the block select devices are maintained in a programmed state having a near depletion mode threshold voltage. NAND strings on more than one layer may be connected to global bit lines on a single layer. By interleaving the NAND strings on each memory level and using two shared bias nodes per block, very little additional overhead is required for the switch devices at each end of the NAND strings. The NAND strings on different memory levels are preferably connected together by way of vertical stacked vias, each preferably connecting to more than one memory level. Each memory level may be produced with less than three masks per level.

Memory adapted to provide dedicated and or shared memory to multiple processors and method therefor

Abstract: Briefly, in accordance with one embodiment of the invention, a portable communication device may have multiple processors and a memory. Portions of the memory may only be accessible by one of the processors.

Shared memory architecture in GPS signal processing

Abstract: A shared memory architecture for a GPS receiver, wherein a processing memory is shared among the different processing functions, such as the correlator signal processing, tracking processing, and other applications processing. The shared memory architecture within the GPS receiver provides the memory necessary for signal processing operations, such as the massively parallel processing, while conserving memory cost by re-using that same memory for other GPS and non-GPS applications. The shared memory architecture for a GPS receiver provided in accordance with the principles of this invention thereby significantly minimize the costly memory requirement often required of extremely fast signal acquisition of a GPS receiver.

Method for dynamically adjusting a memory page closing policy

Abstract: A method for dynamically adjusting a memory page-closing policy for computer systems employing various types of DRAM memory partitioned into one or more memory banks, and circuitry for implementing the method. In general, the method comprises monitoring memory accesses to memory banks and dynamically adjusting the memory page closing policy for those memory bank based on locality characteristics of its memory accesses so that memory latencies are reduced. In one embodiment, in response to memory requests from a computer system processor, memory accesses to the DRAM memory are made on a page-wise basis. As each memory page is accessed, a page-miss, page-hit or page-hit state is produced. Depending on the page access states, which generally will reflect the locality characteristics of (an) application(s) accessing the memory, a page-close set point is adjusted. When a timing count corresponding to the page exceeds the page-close set point, the memory page is closed.

DRAM power management

Abstract: A system and method for limiting power consumption of a computer memory system. The system and method includes selecting a memory access rate. The selected memory access rate corresponds to a desired average memory power consumption rate. A first time interval is started as a current time interval. A memory system is accessed. If the memory access rate has not been exceeded, then the access is applied to the memory system. Alternatively, if the memory access rate has been exceeded, then the access is delayed until the current time interval has expired and a

Semiconductor memory, memory device, and memory card

Abstract: A memory apparatus packaged in one package is provided which includes first data terminals, first address terminals, a status terminal, and memory chips integrated in one semiconductor substrate, one of the memory chips being a nonvolatile memory. Each of the memory chips includes data terminals and address terminals. The data terminals of each of the memory chips are connected to the first data terminals, and the address terminals of each of the memory chips are connected to the first address terminals. The status terminal is arranged to output a status signal which indicates when the nonvolatile memory is in a ready status or in a busy status.

System and method for multi-modal memory controller system operation

Abstract: A memory controller system for processing memory access requests comprising a first memory controller operable to address a first plurality of memory modules a second memory controller operable to address a second plurality of memory modules, the first and second memory controllers configurable to process a memory transaction in an operational mode of the memory controller system selected from the group consisting of an independent cell mode, a multiplexer-mode (mux-mode), and a lockstep mode, and a bus interface block operable to convey the memory transaction to both of the first and second memory controllers is provided.

Methods and apparatus for backing up a memory device

Abstract: A memory module for a computer-based system. The memory module includes at least one memory device that requires periodic refresh signals to maintain data and is mounted on the memory module, and a circuit mounted on the memory module and configured to retain data stored on the memory device when the computer-based system loses power. In a separate embodiment, a control circuit is configured to logically detach at least one memory device from at least one memory controller when a computer-based system loses power, and retain data stored on the memory device when the computer-based system loses power.