Texas Instruments Corporation

Kevin A McGonagle

Tai Huu Nguyen

Harland Glenn Hopkins

Jay Bryan Reimer

Patrick J Smith

Yi Luo

Jason A Jones

Duy Q Nguyen

Dan K Bui

Jason A T Jones

Irvinderpal S Ghai

Kenneth C Kelly

Victor A Liu

A digital signal processing system includes multiple processor subsystems, an external input/output port (XPORT), and an XPORT arbiter. The processor subsystems each include a processor core and a DMA controller. The XPORT arbiter arbitrates between the processor cores and between the DMA controllers, and further arbitrates between processor control or DMA control of the XPORT. Upon a request signal from a DMA controller, the XPORT arbiter asserts a hold signal to the processor cores. The processor cores respond by asserting a hold acknowledge signal. A processor core will delay the hold acknowledge signal until through with the XPORT. The arbiter, then asserts a grant signal to the DMA controller requesting access. The arbiter may assert a grant signal to a processor core requesting access. However, the processor core's access will be stalled as long as the hold signal is asserted.

External bus arbitration technique for multicore DSP device

A digital signal processing chip having a multiple processor cores with corresponding processor subsystems, a shared component, and a clock tree, is disclosed herein. A clock tree distributes clock signals to the processor cores and the shared component. The clock tree can be configured to disable one or more of the processor cores and the shared component by blocking the corresponding clock signal. This may advantageously conserve power. However, the clock tree is configured to preserve the clock signal to the shared component as long as at least one of the processor cores has not disabled the shared component. That is, to block the clock signal to the shared component, each of the processor cores must disable the shared component. The shared component may, for example, be a shared program memory or an arbiter for an external input/output port. The clock tree may include a register and a series of clock gates. Each of the clock gates blocks the clock signal when a gate signal is de-asserted. The gate signals are generated from enablement bits in the register. The clock signal for the shared component is gated by a clock gate that blocks the clock signal only if each of the processor cores have disabled their enablement bit for the shared component.

Unanimous voting for disabling of shared component clocking in a multicore DSP device

A multi-core digital signal processor is disclosed having a shared program memory with conditional write protection. In one embodiment, the digital signal processor includes a shared program memory, an emulation logic module, and multiple processor cores each coupled to the shared program memory by corresponding instruction buses. The emulation logic module preferably determines the operating modes of each of the processors, e.g., whether they are operating in a normal mode or an emulation mode. In the emulation mode, the emulation logic can alter the states of various processor hardware and the contents of various registers and memory. The instruction buses each include a read/write signal that, while their corresponding processor cores are in normal mode, is maintained in a read state. On the other hand, when the processor cores are in the emulation mode, the processor cores are allowed to determine the state of the instruction bus read/write signals. Each instruction bus read/write signal is preferably generated by a logic gate that prevents the processor core from affecting the read/write signal value in normal mode, but allows the processor core to determine the read/write signal value in emulation mode. In this manner, the logic gate prevents write operations to the shared program memory when the emulation logic de-asserts a signal indicative of emulation mode, and allows write operations to the shared program memory when the emulation logic asserts the signal indicative of emulation mode. The logic gate is preferably included in a bus interface module in each processor core.

Multicore DSP device having shared program memory with conditional write protection

A DSP device is disclosed having multiple DMA controllers with global DMA access to all volatile memory resources in the DSP device. In a preferred embodiment, each of the DMA controllers is coupled to each of the memory buses and is configured to control each of the memory buses. A memory bus multiplexer may be coupled between the subsystem memory bus and each of the DMA controllers, and an arbiter may be used to set the memory bus multiplexer so as to allow any one of the DMA controllers to control the memory bus. The memory bus may also be controlled by the host port interface via the memory bus multiplexer. A round-robin arbitration technique is used to provide each of the controllers and the host port interface fair access to the memory bus. This approach may advantageously provide increased flexibility in the use of DMA controllers to transfer data from place to place, with only a minimal increase in complexity.

Multicore DSP device having coupled subsystem memory buses for global DMA access

In a digital processing unit having a plurality of digital signal processors, a first digital signal processor can request a direct transfer of a signal group stored in the memory unit of a second digital signal processor. In order to insure that the second digital signal is active, a control signal is generated by the direct memory access controller of the first digital signal processor. The control signal is applied the directly to the memory access controller of the second digital signal processor. When the second digital signal processor is in an IDLE mode, the control signal activates the second digital signal processor.

Apparatus and method for activation of a digital signal processor in an idle mode for interprocessor transfer of signal groups in a digital signal processing unit

As the digital signal processor has become more flexible, the direct memory access controller has assumed greater computational power to permit the core processing unit to perform its specialized processing without responding to signal transfer requests. Not only does the direct memory access controller control the exchange of signal groups between the memory unit and the core processing unit, but the direct memory access controller is also responsible for the transfer of signal groups within the digital signal processor that originate from the serial port, and the interface unit (the unit that implements the direct transfer of signal groups from the memory unit of one digital signal processor to a second signal processor). The direct memory access controller has programmable channels that permit the signal group source component to be coupled to the signal group destination component. The address unit of the direct memory access unit must be able to accommodate a plurality of addressing modes. An arbitration unit prevents conflict between the components. With respect to the host port interface unit, the exchange of the signal groups with the memory unit as performed by a direct transfer, the direct memory access controller preventing conflicting signal groups transfers.

Apparatus and method for the exchange of signal groups between a plurality of components in a digital signal processor having a direct memory access controller

A multi-core DSP device includes a shared program memory to eliminate redundancy and thereby reduce the size and power consumption of the DSP device. Because each of the program cores typically executes the same software program, memory requirements may be reduced by having multiple processor cores share only a single copy of the software. Accordingly, a program memory couples to each of the processor cores by a corresponding instruction bus. Preferably the program memory services two or more instruction requests in each clock cycle. Data is preferably stored in separate memory arrays local to the processor core subsystems and accessible by the processor cores via a dedicated data bus. In one specific implementation, the program memory includes a wrapper that can perform one memory access in the first half of each clock cycle and a second memory access in the second half of each clock cycle. A designated set of instruction buses is allowed to arbitrate for only the first access, and the remaining instruction buses are allowed to arbitrate for only the second access. In this manner, a reduction in on-board memory requirements and associated power consumption may be advantageously reduced.

Growing community of inventors

Sugar Land, TX, United States of America

Kevin A McGonagle