Abstract: Embodiments of the present invention include an apparatus, method, and system for acoustic modeling. In an embodiment, a speech recognition system is provided. The system includes a processing unit configured to divide a received audio signal into consecutive frames having respective frame vectors, an acoustic processing unit (APU), a data bus that couples the processing unit and the APU. The APU includes a local, non-volatile memory that stores a plurality of senones, a memory buffer coupled to the memory, the acoustic processing unit being configured to load at least one Gaussian probability distribution vector stored in the memory into the memory buffer, and a scoring unit configured to simultaneously compare a plurality of dimensions of a Gaussian probability distribution vector loaded into the memory buffer with respective dimensions of a frame vector received from the processing unit and to output a corresponding score to the processing unit.

Abstract: Embodiments of the present invention include an apparatus, method, and system for acoustic modeling. In an embodiment, a speech recognition system is provided. The system includes a processing unit configured to divide a received audio signal into consecutive frames having respective frame vectors, an acoustic processing unit (APU), a data bus that couples the processing unit and the APU. The APU includes a local, non-volatile memory that stores a plurality of senones, a memory buffer coupled to the memory, the acoustic processing unit being configured to load at least one Gaussian probability distribution vector stored in the memory into the memory buffer, and a scoring unit configured to simultaneously compare a plurality of dimensions of a Gaussian probability distribution vector loaded into the memory buffer with respective dimensions of a frame vector received from the processing unit and to output a corresponding score to the processing unit.

Abstract: A memory buffering system is disclosed that arbitrates bus ownership through an arbitration scheme for memory elements in chain architecture. A unified host memory controller arbitrates bus ownership for transfer to a unified memory buffer and other buffers within the chain architecture. The system is used within a communication system with a bus in chain architectures and parallel architectures.

Abstract: A memory buffering system is disclosed that arbitrates bus ownership through an arbitration scheme for memory elements in chain architecture. A unified host memory controller arbitrates bus ownership for transfer to a unified memory buffer and other buffers within the chain architecture. The system is used within a communication system with a bus in chain architectures and parallel architectures.

Abstract: A voice activation system is provided. The voice activation system includes a first module configured to receive an audio signal and output an activation signal if an energy characteristic of the audio signal satisfies a threshold stored in a memory, a control module configured to enable or disable a third state using a control signal, and a speech recognition engine coupled to the first module and the control module, the speech recognition engine configured to transition between a first state, a second state, and the third state. The speech recognition engine transitions from the first state to the second state in response to the activation signal and in response to the control signal being disabled. The speech recognition engine transitions from the first state to the third state in response to the activation signal and in response to the control signal being enabled.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments for booting an application from multiple memories. An embodiment operates by executing in place from a first memory a first portion of the application, loading a second portion of the application from a second memory, and executing the second portion of the application.

Abstract: A voice activation system is provided. The voice activation system includes a first stage configured to output a first activation signal if at least one energy characteristic of a received audio signal satisfies at least one threshold and a second stage configured to transition from a first state to a second state in response to the first activation signal and, when in the second state, to output a second activation signal if at least a portion of a profile of the audio signal substantially matches at least one predetermined profile.

Abstract: Embodiments are directed to reduced power consumption for memory data transfer at high frequency through synchronized clock signaling. Delay locked loop (DLL) circuits are used to generate the synchronized clock signals. A DLL circuit consumes power as long as it is outputting the synchronized clock signals. A power saving apparatus and method are described wherein the DLL circuit is powered on when memory data access is active, while the DLL circuit is powered down when memory access is idle.

Abstract: Systems and methods embed a random-access non-volatile memory array in a managed-NAND system to execute the boot code or other time-sensitive applications. By embedding this random-access non-volatile memory in the managed-NAND system, either on the memory controller chip or as a separate chip within the managed-NAND system package, an application may be read with fast initial access time, alleviating the slow access time limitations of NAND Flash technology. Depending on the size of the application, the system may be configured to read the whole application content or only a time-critical portion from this embedded random-access non-volatile memory array.

Abstract: According to one exemplary embodiment, a system is provided. The system includes a system bridge, a host module including a central processing unit (CPU) coupled to the system bridge, a scalable serial bus coupled to the system bridge, a client controller coupled to the system bridge via the scalable serial bus, and first and second clients coupled to the client controller. The client controller is configured to interpret a communication received from the CPU and to provide the communication to the first client or to the second client based on a content of the communication. The scalable serial bus is configured such that each of the first and second clients have access to the scalable serial bus during respective time slots. The scalable serial bus is configured to scale at least one of a data transfer rate of the scalable serial bus, a number of channels on the scalable serial bus, or a bus width of the scalable serial bus.

Abstract: An approach is described to overcome the rapid consumption of available flash space when frequently modifying files stored on the flash space. This “differential” sector approach determines the correlation between the new content and the old content, and saves only the “delta” part of the old and the new content to the sectorized memory device. A predetermined threshold can be used to determine whether to use the “differential” sector approach or the fixed sector approach, based on the amount of data change in a given memory access request.

Abstract: According to one exemplary embodiment, a host/client system includes a host module, which includes a CPU coupled to a system bridge. The host/client system further includes at least one client having an integrated interface, where the integrated interface is coupled to the system bridge through a scalable serial bus. The system bridge and the integrated interface enable high bandwidth communication between the CPU and the at least one client through the scalable serial bus, thereby allowing control of bus width between the host module and the client.

Abstract: A memory buffering system is disclosed that arbitrates bus ownership through an arbitration scheme for memory elements in chain architecture. A unified host memory controller arbitrates bus ownership for transfer to a unified memory buffer and other buffers within the chain architecture. The system is used within a communication system with a bus in chain architectures and parallel architectures.

Abstract: A memory buffering system is disclosed that arbitrates bus ownership through an arbitration scheme for memory elements in chain architecture. A unified host memory controller arbitrates bus ownership for transfer to a unified memory buffer and other buffers within the chain architecture. The system is used within a communication system with a bus in chain architectures and parallel architectures.

Abstract: Embodiments are directed to reduced power consumption for memory data transfer at high frequency through synchronized clock signaling. Delay locked loop (DLL) circuits are used to generate the synchronized clock signals. A DLL circuit consumes power as long as it is outputting the synchronized clock signals. A power saving apparatus and method are described wherein the DLL circuit is powered on when memory data access is active, while the DLL circuit is powered down when memory access is idle.

Abstract: A voice activation system is provided. The voice activation system includes a first stage configured to output a first activation signal if at least one energy characteristic of a received audio signal satisfies at least one threshold and a second stage configured to transition from a first state to a second state in response to the first activation signal and, when in the second state, to output a second activation signal if at least a portion of a profile of the audio signal substantially matches at least one predetermined profile.

Abstract: Embodiments of the present invention include an apparatus, method, and system for acoustic modeling. In an embodiment, a speech recognition system is provided. The system includes a processing unit configured to divide a received audio signal into consecutive frames having respective frame vectors, an acoustic processing unit (APU), a data bus that couples the processing unit and the APU. The APU includes a local, non-volatile memory that stores a plurality of senones, a memory buffer coupled to the memory, the acoustic processing unit being configured to load at least one Gaussian probability distribution vector stored in the memory into the memory buffer, and a scoring unit configured to simultaneously compare a plurality of dimensions of a Gaussian probability distribution vector loaded into the memory buffer with respective dimensions of a frame vector received from the processing unit and to output a corresponding score to the processing unit.

Abstract: One embodiment of the present invention relates to a method for communicating NOR-type flash specific memory commands from a DRAM memory controller to a NOR-type flash memory array without disrupting DRAM operation. In this embodiment flash specific commands are channeled from the DRAM controller to the flash device by using the DRAM protocol as a transport layer. Data to be written to the NOR-type flash memory array are loaded into a data register and a sequence of programming commands are loaded into a mode register as a series of mode register write operations. Once the entire sequence of programming commands is loaded the NOR-type flash memory array the data in the data register is loaded into the NOR-type flash memory array. Other methods and circuits are also disclosed.

Abstract: A system is presented that facilitates masking data in write data bound for a memory device. The system contains a memory controller and a memory array communicatively coupled to the memory controller. The memory controller sends a write command and write data to the memory array and the memory array updates data contained therein based upon the write command and write data. If the write operation requires a byte mask, the controller sends a byte mask command via a command bus linking the controller and the memory array. Accordingly, separate and dedicated byte mask pins or bus is not necessary to convey byte mask information.

Abstract: A system is provided that facilitates read access in a memory device. The system comprises a plurality of row addresses buffers that store high order addresses associated with one or more software threads. The system further comprises a plurality of row data buffers. The row data buffers are each associated with at least one row address buffer and store row data within the range of the high order addresses of the row address buffers. The system increase memory device performance by limiting the latency associated with context switching. The plurality of row address buffers and row data buffers enables software threads to associate with one or more buffers and maintain efficient subsequent memory accesses despite context switching.