Vertical and horizontal broadcast of shared operands

A method comprising:

• fetching, via memory interfaces, parameter values from a memory;
• broadcasting, from the memory interfaces, the parameter values to mutually exclusive subsets of rows and columns of processor element arrays;
• and
• performing, at the processor element arrays, operations on the parameter values,
• wherein subsets of the processor element arrays in each row are implemented in corresponding single-instruction-multiple-data (SIMD) units, and broadcasting the parameter values comprises broadcasting the parameter values to the SIMD units comprising the processor element arrays in rows associated with the memory interfaces, and
• wherein subsets of the SIMD units are implemented in corresponding workgroup processors, and broadcasting the parameter values comprises broadcasting the parameter values to columns of the processor element arrays that are implemented across the SIMD units in the workgroup processors.

A non-transitory computer readable medium embodying a set of executable instructions, the set of executable instructions to manipulate a computer system to perform a portion of a process to fabricate at least part of a processor, the processor comprising:

• processor element arrays distributed in rows and columns, wherein the processor element arrays are configured to perform operations on parameter values;
• and
• memory interfaces configured to broadcast sets of the parameter values to mutually exclusive subsets of the rows and columns of the processor element arrays;
• single-instruction-multiple-data (SIMD) units that comprise subsets of the processor element arrays in corresponding rows, and
• wherein the memory interfaces broadcast the parameter values to the SIMD units comprising the processor element arrays in rows associated with the memory interfaces;
• workgroup processors comprising subsets of the SIMD units, and
• wherein the memory interfaces broadcast the parameter values to columns of processor element arrays that are implemented across the SIMD units in the workgroup processors;
• and
• a memory fabric configured to interconnect with an external memory that stores the parameter values, and
• wherein the memory interfaces are configured to access the parameter values from the external memory via the memory fabric.