A renaming unit configured to rename source operands of instructions in a group. A renaming register maintains architectural to physical register mappings. Architectural to physical register mappings propagate from the renaming register through a chain of update units (U) over bus lines denoted with the architectural registers 0 to L. Update units (U) sequentially, in program order, insert physical register identifiers PR(i) allocated to instructions I(i) with destination operands DOP(i) on bus lines denoted with the destination operands DOP(i). Source operands of an instruction I(i) may be renamed to physical register identifiers after physical register identifiers allocated to instructions older than I(i) are sequentially, in program order, inserted on the bus lines, but before physical register identifiers allocated to I(i) and younger instructions are inserted on the bus lines. A source operand SOP(i) is renamed to a physical register identifier that propagates on a bus line denoted with SOP(i).

Throttling recommendations for a systolic array may be arbitrated. Throttling recommendations may be received at an arbiter for a systolic array from different sources, such as one or more monitors implemented in an integrated circuit along with the systolic array or sources external to the integrated circuit with the systolic array. A strongest throttling recommendation may be selected. The rate at which data enters the systolic array may be modified according to the strongest throttling recommendation.

A computer system, processor, programming instructions and/or method of processing data that includes a main register file having a plurality of entries for storing data; an accumulator register file having a plurality of entries for storing data wherein multiple main register file entries are mapped to one accumulator register file entry in the at least one accumulator register file; a logical register mapper to track and map logical registers to main register file entries, and a history buffer. Processing wide data width instructions includes evicting and restoring information from a single primary entry in the logical register mapper through a single read or write port in the logical register mapper without evicting or restoring the remaining other multiple main register file entries mapped in the accumulator register.

Technologies disclosed herein provide cryptographic computing with cryptographically encoded pointers in multi-tenant environments. An example method comprises executing, by a trusted runtime, first instructions to generate a first address key for a private memory region in the memory and generate a first cryptographically encoded pointer to the private memory region in the memory. Generating the first cryptographically encoded pointer includes storing first context information associated with the private memory region in first bits of the first cryptographically encoded pointer and performing a cryptographic algorithm on a slice of a first linear address of the private memory region based, at least in part, on the first address key and a first tweak, the first tweak including the first context information. The method further includes permitting a first tenant in the multi-tenant environment to access the first address key and the first cryptographically encoded pointer to the private memory region.

A data processing apparatus, a method of operating a data processing apparatus, a non-transitory computer readable storage medium, and an instruction are provided. The instruction specifies a first source register, a second source register, and a set of N accumulation registers. In response to the instruction control signals are generated, causing processing circuitry to extract N data elements from content of the first source register, perform a multiplication of each of the N data elements by content of the second source register, and apply a result of each multiplication to content of a respective target register of the set of N accumulation registers. As a result plural (N) multiplications are performed in a manner that effectively provides a multiplier N times the register width, but without requiring the register file to be made N times larger.

Embodiments are directed to immutable watermarking for authenticating and verifying artificial intelligence (AI)-generated output. An embodiment of a system includes a hardware accelerator to perform processing related to a machine learning (ML) model and one or more processors including a hash generator. In one implementation, the hash generator is to identify a global unique identifier (GUID) for the ML model, generate a digital signature for content generated by an inference stage of the ML model, the digital signature based on at least the GUID of the ML model and the content generated by the ML model, and transmit the content and the digital signature to a content consumer platform.

Systems and methods for updating metadata. In some embodiments, in response to detecting an instruction executed by a hardware system, a source location of the instruction may be identified. First metadata associated with the instruction may be used to determine whether the instruction is allowed. In response to determining that the instruction is allowed, the source location of the instruction may be associated with second metadata.

Data processing circuitry comprises out-of-order instruction execution circuitry; register mapping circuitry to map zero or more architectural processor registers relating to execution of that program instruction to respective ones of a set of physical processor registers; commit circuitry to commit, in a program code order, the results of executed program instructions, the commit circuitry being configured to access a data store which stores register tag data to indicate which physical registers mapped by the register mapping circuitry relate to a given program instruction; fault detection circuitry to detect a memory access fault in respect of a vector memory access operation and to generate fault indication data indicative of an element earliest in the element order for which a memory access fault was detected; a fault indication register to store the fault indication data, in which the register mapping circuitry is configured to generate a register mapping for a program instruction for any architectural processor registers relating to execution of that program instruction other than the fault indication register; and control circuitry to encode the fault indication data, applicable to a program instruction not yet committed by the commit circuitry, to register tag data associated with that program instruction.

The problem to be solved is to seek an alternative to known addressing methods which provides the same or similar effects or is more secure. Solution The problem is solved by a method (40) of addressing memory in a data-processing apparatus (10) comprising, when a central processing unit (11), while performing a task (31, 32, 33, 34) of the apparatus (10), executes an instruction involving a pointer (59) into a segment (s, r, d, h, f, o, i, c) of the memory: decoding the instruction by means of an instruction decoder (12), generating a virtual address (45) within the memory by means of a safe pointer operator (41) operating on the pointer (59), augmenting the virtual address (45) by an identifier (43) of the task (31, 32, 33, 34) and an identifier (44) of the segment (s, r, d, h, f, o, i, c), said identifiers (43, 44) being hardware-controlled (42), and, based on the augmented address (45), dereferencing the pointer (59) via a memory management unit (13).

Provided is a neural network accelerator which performs a calculation of a neural network provided with layers, the neural network accelerator including a kernel memory configured to store kernel data related to a filter, a feature map memory configured to store feature map data which are outputs of the layers, and a Processing Element (PE) array including PEs arranged along first and second directions, wherein each of the PEs performs a calculation using the feature map data transmitted in the first direction from the feature map memory and the kernel data transmitted in the second direction from the kernel memory, and transmits a calculation result to the feature map memory in a third direction opposite to the first direction.