Aspects for vector operations in neural network are described herein. The aspects may include a vector caching unit configured to store a first vector and a second vector, wherein the first vector includes one or more first elements and the second vector includes one or more second elements. The aspects may further include one or more adders and a combiner. The one or more adders may be configured to respectively add each of the first elements to a corresponding one of the second elements to generate one or more addition results. The combiner may be configured to combine a combiner configured to combine the one or more addition results into an output vector.

An apparatus and method for performing vector index loads and stores. For example, one embodiment of a processor comprises: a vector index register to store a plurality of index values; a mask register to store a plurality of mask bits; a vector register to store a plurality of vector data elements loaded from memory; and vector index load logic to identify an index stored in the vector index register to be used for a load operation using an immediate value and to responsively combine the index with a base memory address to determine a memory address for the load operation, the vector index load logic to load vector data elements from the memory address to the vector register in accordance with the plurality of mask bits.

In an embodiment, a coprocessor may include a bypass indication which identifies execution circuitry that is not used by a given processor instruction, and thus may be bypassed. The corresponding circuitry may be disabled during execution, preventing evaluation when the output of the circuitry will not be used for the instruction. In another embodiment, the coprocessor may implement a grid of processing elements in rows and columns, where a given coprocessor instruction may specify an operation that causes up to all of the processing elements to operate on vectors of input operands to produce results. Implementations of the coprocessor may implement a portion of the processing elements. The coprocessor control circuitry may be designed to operate with the full grid or partial grid, reissuing instructions in the partial grid case to perform the requested operation. In still another embodiment, the coprocessor may be able to fuse vector mode operations.

This invention discloses a novel paradigm, method and apparatus for Matrix Computing which include a novel machine architecture with an embedded storage space for holding matrices and arrays for computing which can be accessed by its columns or by its rows or both concurrently. A large capacity multi length instruction set with instructions and methods to load, store and compute with these matrices and arrays are also disclosed; a method and apparatus to secure, share, lock and unlock this embedded space for matrices under the control of an Operating System or a Virtual Machine Monitor by a plurality of threads and processes are also disclosed. A novel method and apparatus to handle immediate operands used by Immediate Instructions are also disclosed. The structure of the instructions with some key fields and a method for determining instruction length easily are also disclosed.

Vector processing engines (VPEs) employing a tapped-delay line(s) for providing precision filter vector processing operations with reduced sample re-fetching and power consumption are disclosed. Related vector processor systems and methods are also disclosed. The VPEs are configured to provide filter vector processing operations. To minimize re-fetching of input vector data samples from memory to reduce power consumption, a tapped-delay line(s) is included in the data flow paths between a vector data file and execution units in the VPE. The tapped-delay line(s) is configured to receive and provide input vector data sample sets to execution units for performing filter vector processing operations. The tapped-delay line(s) is also configured to shift the input vector data sample set for filter delay taps and provide the shifted input vector data sample set to execution units, so the shifted input vector data sample set does not have to be re-fetched during filter vector processing operations.

A vector processor is disclosed including a variety of variable-length instructions. Computer-implemented methods are disclosed for efficiently carrying out a variety of operations in a time-conscious, memory-efficient, and power-efficient manner. Methods for more efficiently managing a buffer by controlling the threshold based on the length of delay line instructions are disclosed. Methods for disposing multi-type and multi-size operations in hardware are disclosed. Methods for condensing look-up tables are disclosed. Methods for in-line alteration of variables are disclosed.

A Vector Generate Mask instruction. For each element in the first operand, a bit mask is generated. The mask includes bits set to a selected value starting at a position specified by a first field of the instruction and ending at a position specified by a second field of the instruction.

Methods, systems, and articles of manufacture for merging and sorting arrays on a processor are provided herein. A method includes splitting an input array into multiple sub-arrays across multiple processing elements; merging the multiple sub-arrays into multiple vectors; and sorting the multiple vectors by comparing and swapping one or more vector elements among the multiple vectors.

In an embodiment, a coprocessor may include a bypass indication which identifies execution circuitry that is not used by a given processor instruction, and thus may be bypassed. The corresponding circuitry may be disabled during execution, preventing evaluation when the output of the circuitry will not be used for the instruction. In another embodiment, the coprocessor may implement a grid of processing elements in rows and columns, where a given coprocessor instruction may specify an operation that causes up to all of the processing elements to operate on vectors of input operands to produce results. Implementations of the coprocessor may implement a portion of the processing elements. The coprocessor control circuitry may be designed to operate with the full grid or partial grid, reissuing instructions in the partial grid case to perform the requested operation. In still another embodiment, the coprocessor may be able to fuse vector mode operations.

A reconfigurable vector processor is described that allows the size of its vector units to be changed in order to process vectors of different sizes. The reconfigurable vector processor comprises a plurality of processor units. Each of the processor units comprises a control unit for decoding instructions and generating control signals, a scalar unit for processing instructions on scalar data, and a vector unit for processing instructions on vector data under control of control signals. The reconfigurable vector processor architecture also comprises a vector control selector for selectively providing control signals generated by one processor unit of the plurality of processor units to the vector unit of a different processor unit of the plurality of processor units.