US 10,242,426 B2 · Multistage collector for outputs in multiprocessor systems

Abstract

Aspects include a multistage collector to receive outputs from plural processing elements. Processing elements may comprise (each or collectively) a plurality of clusters, with one or more ALUs that may perform SIMD operations on a data vector and produce outputs according to the instruction stream being used to configure the ALU(s). The multistage collector includes substituent components each with at least one input queue, a memory, a packing unit, and an output queue; these components can be sized to process groups of input elements of a given size, and can have multiple input queues and a single output queue. Some components couple to receive outputs from the ALUs and others receive outputs from other components. Ultimately, the multistage collector can output groupings of input elements. Each grouping of elements (e.g., at input queues, or stored in the memories of component) can be formed based on matching of index elements.

Claims14

Dependent claims

A method of parallelized computation in a distributed multiprocessor system having a plurality of computation units, comprising:

executing a plurality of programs in said plurality of computation units to determine respective pairings of data elements and code modules to be executed subsequently in the multiprocessor system and outputting the pairings from a plurality of ports of said computation units, wherein one or more of the data elements comprises computer graphics ray tracing information identifying a shape, and identifiers for rays to be tested for intersection with the shape;
collecting the data elements of the pairings that reference the same code module by extracting each of the data elements from its original pairing and grouping the extracted data elements referencing the same code module into larger groupings, each grouping maintaining an association with its respective code module;
terminating the collecting of data elements for a specific grouping based on a heuristic;
and
submitting to the multiprocessor system, the specific grouping of data elements for use during execution of the code module associated with that specific grouping of data elements by one or more of said computation units.

2 more elements

A non-transitory computer readable medium having stored thereon computer executable instructions that when executed cause at least one processor to:

execute a plurality of programs in a plurality of computation units of a multiprocessor system to determine respective pairings of data elements and code modules to be executed subsequently in said multiprocessor system and output the pairings from a plurality of ports of said computation units;
collect the data elements of the pairings that reference the same code module by extracting each of the data elements from its original pairing and group the extracted data elements referencing the same code module into larger groupings, each grouping maintaining an association with its respective code module, wherein one or more of the data elements comprises computer graphics ray tracing information identifying a shape, and identifiers for rays to be tested for intersection with the shape;
terminate the collecting of data elements for a specific grouping based on a heuristic;
and
submit to the multiprocessor system, the specific grouping of data elements for use during execution of the code module associated with that specific grouping of data elements by one or more of said computation units.

2 more elements

Dependencies are inferred from the claim text.

Multistage collector for outputs in multiprocessor systems

Life of this patent

Multistage collector for outputs in multiprocessor systems

Life of this patent

Abstract

Claims14