Mill Computing, Inc.

Mil Computing, Inc.

Sebastien Paul Maurice Mirolo

David Arthur Yost

Roger Rawson Godard

Arthur David Kahlich

Nachum M Kanovsky

A computer processor is provided that employs a plurality of operand storage elements that store operand data values and associated meta-data as unitary operand data elements as well as at least one functional unit that performs operations that produce and access the unitary operand data elements stored in the plurality of operand storage elements. The meta-data associated with a given operand data value as part of a unitary operand data element can specify type of the unitary operand data element (e.g., vector or scalar), elemental width and floating-point error flags. The meta-data can also be used to define special operand data values (e.g., Not-a-Result and None). The meta-data is useful in optimizing execution, such as in speculation and vectorized SIMD operations. The computer processor can also support a number of particular vector operations that are useful in optimizing execution of vectorized SIMD operations.

Computer processor employing operand data with associated meta-data

A computer processor including a first memory structure that operates over multiple cycles to temporarily store operands referenced by at least one instruction. A plurality of functional units performs operations that produce and access operands stored in the first memory structure. A second memory structure is provided, separate from the first memory structure. The second memory structure is configured as a dedicated memory for storage of operands copied from the first memory structure. The second memory structure is organized with a byte-addressable memory space and each operand stored in the second memory structure is accessed by a given byte address into the byte-addressable memory space.

Computer processor employing byte-addressable dedicated memory for operand storage

A computer processor including a plurality of functional units that performs operations that produce result operands at different characteristic latencies over multiple cycles. An interconnect network provides data paths for transfer of operand data between functional units. The interconnect network includes first and second crossbar parts. The first crossbar part is configured to route result operands produced with the lowest characteristic latency to any other functional unit. The second crossbar part is configured to route result operands with higher characteristic latency relative to the lowest characteristic latency to the first crossbar part where such result operands are in turn routed to any functional unit. In another aspect, the functional units can be organized as multiple slots where each slot can produce multiple result operands of different characteristic latencies in the same cycle, and wherein each slot employs separate result registers for each characteristic latency present on the slot.

Computer processor employing split crossbar circuit for operand routing and slot-based organization of functional units

A computer processor and corresponding method of operation employs execution logic that includes at least one functional unit and operand storage that stores data that is produced and consumed by the at least one functional unit. The at least one functional unit is configured to execute a deferred operation whose execution produces result data. The execution logic further includes a retire station that is configured to store and retire the result data of the deferred operation in order to store such result data in the operand storage, wherein the retire of such result data occurs at a machine cycle following issue of the deferred operation as controlled by statically-assigned parameter data included in the encoding of the deferred operation.

Computer processor with deferred operations

A computer processor including a plurality of storage elements logically organized as a fixed length queue referenced by logical temporal addresses. The fixed length queue operates over multiple cycles to temporarily store operands referenced by at least one instruction utilizing the logical temporal addresses. A plurality of functional units performs operations over the multiple cycles, wherein the operations produce and access operands stored in the logical fixed length queue. Operands can be added to the front of the logical fixed length queue according to the temporal order that operands are produced by the functional units, and operands can drop from the end of the logical fixed length queue as operands are added to the front of the fixed length queue. A plurality of operands produced by the plurality of functional units (possibly with different latencies in producing such operands) can be added to the logical fixed length queue in a single cycle. A plurality of operands operated on by the functional units can be accessed from the logical fixed length queue in a single cycle.

Growing community of inventors

San Francisco, CA, United States of America

Sebastien Paul Maurice Mirolo