Arm Limited

Michael Alexander Kennedy

Anthony Jebson

Neil Burgess

Andrew John Turner

Simon John Craske

Chiloda Ashan Senarath Pathirane

Zichao Xie

Richard Anthony Lane

Karel Hubertus Gerardus Walters

Richard Roy Grisenthwaite

Ian Michael Caulfield

Martin Weidmann

Timothy Nicholas Hay

Jack William Derek Andrew

Tai Li

A multiplier circuit is provided to multiply a first operand and a second operand. The multiplier circuit includes a carry-save adder network comprising a plurality of carry-save adders to perform partial product additions to reduce a plurality of partial products to a redundant result value that represents a product of the first operand and the second operand. A number of the carry-save adders that is used to generate the redundant result value is controllable and is dependent on a width of at least one of the first operand and the second operand.

Controlling carry-save adders in multiplication

An interrupt controller comprises issue circuitry to issue interrupt requests to a processing element and control circuitry to detect presence of a race condition in association with at least one pending interrupt request to be issued, and to set a barrier indicator when the race condition has been resolved. In response to the race condition being present, the issue circuitry is configured to select one of the at least one pending interrupt requests, to issue to the processing element the selected pending interrupt request followed by a dummy request over a path that ensures that the processing element receives the selected pending interrupt request prior to receiving the dummy request. On receiving an acknowledgement indicating that the processing element has received the dummy request, the control circuitry is then configured to set the barrier indicator.

Interrupt controller

A multiplier circuit is described in which sub-products calculated in a first stage of a carry-save adder (CSA) network are output early, processed by applying a processing function, and re-injected into a subsequent stage of the CSA network to add the processed sub-products. This allows a CSA network provided for multiplication operations to be reused for operations which require sub-products to be processed and added, such as floating-point dot product operations performed on floating-point values represented in bfloatl6 format.

Multiplier circuit

An apparatus comprises instruction decoding circuitry to generate micro-operations in response to program instructions; and processing circuitry to perform data processing in response to the micro-operations generated by the instruction decoding circuitry. In response to a predicated vector instruction, the instruction decoding circuitry reads or predicts an estimated value of the predicate value, and depending on the estimated value, varies a composition of at least one micro-operation generated in response to the predicated vector instruction. This can enable more efficient use of hardware resources in the processing circuitry.

Varying micro-operation composition based on estimated value of predicate value for predicated vector instruction

A configurable SIMD multiplication circuit is provided to perform multiplication on a multiplicand operand M and multiplier operand R with varying data element sizes supported. For each result element generated based on corresponding elements of the multiplicand operand M and the multiplier operand R, the multiplication is performed according to radix-N modified Booth multiplication, where N=2and P&#x2265;3. A Booth digit selection scheme is described for improving the efficiency with which higher radix modified Booth multiplication can be implemented in a configurable SIMD multiplier.

Configurable SIMD multiplication circuit

An apparatus and method are provided for performing multiply-and-accumulate-products (MAP) operations. The apparatus has processing circuitry for performing data processing, the processing circuitry including an adder array having a plurality of adders for accumulating partial products produced from input operands. An instruction decoder is provided that is responsive to a MAP instruction specifying a first J-bit operand and a second K-bit operand, to control the processing circuitry to enable performance of a number of MAP operations, where the number is dependent on a parameter. For each performed MAP operation, the processing circuitry is arranged to generate a corresponding result element representing a sum of respective E&#xd7;F products of E-bit portions within an X-bit segment of the first operand with F-bit portions within a Y-bit segment of the second operand, where E&#x3c;X&#x2264;J and F&#x3c;Y&#x2264;K. In response to the MAP instruction, the instruction decoder is configured to control the processing circuitry to perform a rearrangement operation to rearrange the portions of at least one of the first operand and the second operand to form transformed first and second operands so that the E&#xd7;F products that are required to be summed are aligned with a number of columns of adders within the adder array. Further, the adder array is controlled in dependence on the transformed first and second operands to add the required E&#xd7;F products using the number of columns of adders within the adder array. Further, the rearrangement performed by the rearrangement operation is controlled in dependence on the parameter such that at least one column of adders used when the parameter indicates the first number of MAP operations is reused when the parameter indicates a second number of MAP operations different to the first number. This provides a particularly area and power efficient implementation for handling MAP operations.

Apparatus and method for performing multiply-and-accumulate-products operations

A data processing system includes an interrupt controller having a priority level arbitrator and trigger circuitry. The priority level arbitrator and the trigger circuitry operate in parallel to process interrupt signals received by an interrupt signal receiver. The trigger circuitry generates a trigger signal initiating interrupt processing before the priority level arbitrator has completed its arbitration determination at an arbitration-completed time. If the interrupt processing triggered by the trigger signal was inappropriate, then is terminated once the result of the arbitration is known after the arbitration-completed time.

Speculative interrupt signalling

A data processing apparatus has a plurality of processors and a plurality of interrupt interfaces each for handling interrupt requests from a corresponding processor. An interrupt distributor controls routing of interrupt requests to the interrupt interfaces. A shared interrupt request is serviceable by multiple processors. In response to the shared interrupt request, a target interrupt interface issues an interrupt ownership request to the interrupt distributor, without passing the shared interrupt request to the corresponding processor, if it estimates that the corresponding processor is available for servicing the shared interrupt request. The shared interrupt request is passed to the corresponding processor when an ownership confirmation is received from the interrupt distributor indicating that the processor has been selected for servicing the shared interrupt request.

Handling interrupts in a multi-processor system

An interrupt controller includes a priority level arbitrator () including multiple stages. The stages include at least one stage comprising a plurality of interrupt selectors formed of a multiplexer () for selecting between a pair of potentially concurrently asserted interrupt signals in dependence upon selection data. The selection data is determined in advance by a priority level comparator () using priority level data associated with the respective interrupt signals.

Interrupt signal arbitration

A data processing device comprises a plurality of system registers and a set of interrupt handling registers for controlling handling of an incoming interrupt. The device also includes processing circuitry configured to execute software of the plurality of execution levels, and interrupt controller circuitry configured to route said incoming interrupts to interrupt handling software that is configured to run at one of said plurality of execution levels, and register access control circuitry configured to dynamically control access to at least some of said interrupt handling registers in dependence upon one of said plurality of execution levels that said incoming interrupt is routed to. The interrupt handling software configured to run at a particular execution level does not have access to interrupt handling registers for handling a different incoming interrupt that is routed to interrupt handling software that is configured to run at a more privileged execution level.

Growing community of inventors

Cambridge, United Kingdom

Michael Alexander Kennedy