Abstract: A cache of a processor includes a cache controller to implement a cache management policy for the insertion and replacement of cache lines of the cache. The cache management policy assigns replacement priority levels to each cache line of at least a subset of cache lines in a region of the cache based on a comparison of a number of accesses to a cache set having a way that stores a cache line since the cache line was last accessed to a reuse distance determined for the region of the cache, wherein the reuse distance represents an average number of accesses to a given cache set of the region between accesses to any given cache line of the cache set.

Abstract: Systems, apparatuses, and methods for cache management based on access type priority are disclosed. A system includes at least a processor and a cache. During a program execution phase, certain access types are more likely to cause demand hits in the cache than others. Demand hits are load and store hits to the cache. A run-time profiling mechanism is employed to find which access types are more likely to cause demand hits. Based on the profiling results, the cache lines that will likely be accessed in the future are retained based on their most recent access type. The goal is to increase demand hits and thereby improve system performance. An efficient cache replacement policy can potentially reduce redundant data movement, thereby improving system performance and reducing energy consumption.

Abstract: An electronic device includes a controller functional block and a computational functional block. During operation, while the computational functional block executes a test portion of a workload at at least one precision level, the controller functional block monitors a behavior of the computational functional block. Based on the behavior of the computational functional block while executing the test portion of the workload at the at least one precision level, the controller functional block selects a given precision level from among a set of two or more precision levels at which the computational functional block is to execute a remaining portion of the workload. The controller functional block then configures the computational block to execute the remaining portion of the workload at the given precision level.

Abstract: A technique for processing qubits in a quantum computing device is provided. The technique includes determining that, in a first cycle, a first quantum processing region is to perform a first quantum operation that does not use a qubit that is stored in the first quantum processing region, identifying a second quantum processing region that is to perform a second quantum operation at a second cycle that is later than the first cycle, wherein the second quantum operation uses the qubit, determining that between the first cycle and the second cycle, no quantum operations are performed in the second quantum processing region, and moving the qubit from the first quantum processing region to the second quantum processing region.

Abstract: Systems, apparatuses, and methods for implementing cache line re-reference interval prediction using a physical page address are disclosed. When a cache line is accessed, a controller retrieves a re-reference interval counter value associated with the line. If the counter is less than a first threshold, then the address of the cache line is stored in a small re-use page buffer. If the counter is greater than a second threshold, then the address is stored in a large re-use page buffer. When a new cache line is inserted in the cache, if its address is stored in the small re-use page buffer, then the controller assigns a high priority to the line to cause it to remain in the cache to be re-used. If a match is found in the large re-use page buffer, then the controller assigns a low priority to the line to bias it towards eviction.

Abstract: Systems, apparatuses, and methods for cache management based on access type priority are disclosed. A system includes at least a processor and a cache. During a program execution phase, certain access types are more likely to cause demand hits in the cache than others. Demand hits are load and store hits to the cache. A run-time profiling mechanism is employed to find which access types are more likely to cause demand hits. Based on the profiling results, the cache lines that will likely be accessed in the future are retained based on their most recent access type. The goal is to increase demand hits and thereby improve system performance. An efficient cache replacement policy can potentially reduce redundant data movement, thereby improving system performance and reducing energy consumption.

Abstract: A cache of a processor includes a cache controller to implement a cache management policy for the insertion and replacement of cache lines of the cache. The cache management policy assigns replacement priority levels to each cache line of at least a subset of cache lines in a region of the cache based on a comparison of a number of accesses to a cache set having a way that stores a cache line since the cache line was last accessed to a reuse distance determined for the region of the cache, wherein the reuse distance represents an average number of accesses to a given cache set of the region between accesses to any given cache line of the cache set.

Abstract: A method includes receiving, from an origin computing node, a first communication addressed to multiple destination computing nodes in a processor interconnect fabric, measuring a first set of one or more communication metrics associated with a transmission path to one or more of the multiple destination computing nodes, and for each of the destination computing nodes, based on the set of communication metrics, selecting between a multicast transmission mode and unicast transmission mode as a transmission mode for transmitting the first communication to the destination computing node.

Abstract: An electronic device handles memory access requests for data in a memory. The electronic device includes a memory controller for the memory, a last-level cache memory, a request generator, and a predictor. The predictor determines a likelihood that a cache memory access request for data at a given address will hit in the last-level cache memory. Based on the likelihood, the predictor determines: whether a memory access request is to be sent by the request generator to the memory controller for the data in parallel with the cache memory access request being resolved in the last-level cache memory, and, when the memory access request is to be sent, a type of memory access request that is to be sent. When the memory access request is to be sent, the predictor causes the request generator to send a memory request of the type to the memory controller.

Abstract: An electronic device handles memory access requests for data in a memory. The electronic device includes a memory controller for the memory, a last-level cache memory, a request generator, and a predictor. The predictor determines a likelihood that a cache memory access request for data at a given address will hit in the last-level cache memory. Based on the likelihood, the predictor determines: whether a memory access request is to be sent by the request generator to the memory controller for the data in parallel with the cache memory access request being resolved in the last-level cache memory, and, when the memory access request is to be sent, a type of memory access request that is to be sent. When the memory access request is to be sent, the predictor causes the request generator to send a memory request of the type to the memory controller.

Abstract: A method includes receiving, from an origin computing node, a first communication addressed to multiple destination computing nodes in a processor interconnect fabric, measuring a first set of one or more communication metrics associated with a transmission path to one or more of the multiple destination computing nodes, and for each of the destination computing nodes, based on the set of communication metrics, selecting between a multicast transmission mode and unicast transmission mode as a transmission mode for transmitting the first communication to the destination computing node.

Abstract: The deep Q learning technique trains weights of an artificial neural network using a number of unique features, including separate target and prediction networks, random experience replay to avoid issues with temporally correlated training samples, and others. A hardware architecture is described that is tuned to perform deep Q learning. Inference cores use a prediction network to determine an action to apply to an environment. A replay memory stores the results of the action. Training cores use a loss function derived from outputs from both the target and prediction networks to update weights of the prediction neural networks. A high speed copy engine periodically copies weights from the prediction neural network to the target neural network.

Abstract: An electronic device includes a controller functional block and a computational functional block. During operation, while the computational functional block executes a test portion of a workload at at least one precision level, the controller functional block monitors a behavior of the computational functional block. Based on the behavior of the computational functional block while executing the test portion of the workload at the at least one precision level, the controller functional block selects a given precision level from among a set of two or more precision levels at which the computational functional block is to execute a remaining portion of the workload. The controller functional block then configures the computational block to execute the remaining portion of the workload at the given precision level.

Abstract: Systems, methods, and devices for determining a derived counter value based on a hardware performance counter. Example devices include input circuitry configured to input a hardware performance counter value; counter engine circuitry configured to determine the derived counter value by applying a model to the hardware performance counter value; and output circuitry configured to communicate the derived counter value to a consumer. In some examples, the consumer includes an operating system scheduler, a memory controller, a power manager, or a data prefetcher, or a cache controller. In some examples, the processor includes circuitry configured to dynamically change the model during operation of the processor. In some examples, the model includes or is generated by an artificial neural network (ANN).

Abstract: The described embodiments include an apparatus that controls voltages for an integrated circuit chip having a set of circuits. The apparatus includes a switching voltage regulator separate from the integrated circuit chip and two or more low dropout (LDO) regulators fabricated on the integrated circuit chip. The switching voltage regulator provides an output voltage that is received as an input voltage by each of the two or more LDO regulators, and each of the two or more LDO regulators provides a local output voltage, each local output voltage received as a local input voltage by a different subset of the circuits in the set of circuits. During operation, a controller sets an operating point for each of the subsets of circuits based on a combined power efficiency for the subsets of the circuits and the LDO regulators, each operating point including a corresponding frequency and voltage.

Abstract: The described embodiments include an apparatus that controls voltages for an integrated circuit chip having a set of circuits. The apparatus includes a switching voltage regulator separate from the integrated circuit chip and two or more low dropout (LDO) regulators fabricated on the integrated circuit chip. The switching voltage regulator provides an output voltage that is received as an input voltage by each of the two or more LDO regulators, and each of the two or more LDO regulators provides a local output voltage, each local output voltage received as a local input voltage by a different subset of the circuits in the set of circuits. During operation, a controller sets an operating point for each of the subsets of circuits based on a combined power efficiency for the subsets of the circuits and the LDO regulators, each operating point including a corresponding frequency and voltage.

Abstract: The described embodiments include an apparatus that controls voltages for an integrated circuit chip having a set of circuits. The apparatus includes a switching voltage regulator separate from the integrated circuit chip and two or more low dropout (LDO) regulators fabricated on the integrated circuit chip. The switching voltage regulator provides an output voltage that is received as an input voltage by each of the two or more LDO regulators, and each of the two or more LDO regulators provides a local output voltage, each local output voltage received as a local input voltage by a different subset of the circuits in the set of circuits. During operation, a controller sets an operating point for each of the subsets of circuits based on a combined power efficiency for the subsets of the circuits and the LDO regulators, each operating point including a corresponding frequency and voltage.

Abstract: A method and apparatus for transmitting data includes determining whether to apply a mask to a cache line that includes a first type of data and a second type of data for transmission based upon a first criteria. The second type of data is filtered from the cache line, and the first type of data along with an identifier of the applied mask is transmitted. The first type of data and the identifier is received, and the second type of data is combined with the first type of data to recreate the cache line based upon the received identifier.

Abstract: A method and apparatus for transmitting data includes determining whether to apply a mask to a cache line that includes a first type of data and a second type of data for transmission based upon a first criteria. The second type of data is filtered from the cache line, and the first type of data along with an identifier of the applied mask is transmitted. The first type of data and the identifier is received, and the second type of data is combined with the first type of data to recreate the cache line based upon the received identifier.