Abstract: The disclosed semiconductor package includes a first chiplet area for receiving a first chiplet, a second chiplet area for receiving a second chiplet, and a host die coupled to the first and second chiplet areas. The semiconductor package also includes an interconnect directly coupling the first chiplet area to the second chiplet area. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computing device can include at least one memory of a particular type having a plurality of memory channels, and at least one memory of at least one other type having a plurality of links. The computing device can also include remapping circuitry configured to homogenously interleave the plurality of memory channels with the at least one memory of the at least one other type. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: Permute instructions for register-based lookups is described. In accordance with the described techniques, a processor is configured to perform a register-based lookup by retrieving a first result from a first lookup table based on a subset of bits included in an index of a destination register, retrieving a second result from a second lookup table based on the subset of bits included in the index of the destination register, selecting the first result or the second result based on a bit in the index of the destination register that is excluded from the subset of bits, and overwriting data included in the index of the destination register using a selected one of the first result or the second result.

Abstract: A static random-access memory (SRAM) circuit includes an SRAM bitcell coupled to a word line, a bit line and a complementary bit line. A precharge circuit is coupled to the bit line and the complementary bit line and includes a precharge input. A first keeper transistor is coupled to the bit line and a second keeper transistor is coupled to the complementary bit line. A write driver circuit includes a select input receiving a select signal, a write data input, and a write data compliment input, and is operable to write a data bit to the SRAM bitcell. A combinatorial logic circuit provides a precharge signal to the precharge circuit based on the select signal and a bit line precharge signal.

Abstract: Adaptive scheduling of memory requests and processing-in-memory requests is described. In accordance with the described techniques, a memory controller receives a plurality of processing-in-memory requests and a plurality of non-processing-in-memory requests from a host. The memory controller schedules an order of execution for the plurality of processing-in-memory requests and the plurality of non-processing-in-memory requests based at least in part on a processing-in-memory request stall threshold and a non-processing-in-memory request stall threshold. In response to a system switching (e.g., from executing processing-in-memory requests to executing non-processing-in-memory requests or from executing non-processing-in-memory requests to executing processing-in-memory requests), the memory controller modifies the processing-in-memory request stall threshold and the non-processing-in-memory request stall threshold.

Abstract: Automated memory overclocking is described. In accordance with the described techniques, one or more sets of overclocked memory settings of a memory are automatically selected for performance testing and stability testing of the memory. The one or more sets of the overclocked memory settings are tested for performance of the memory and a performance indication is output for each of the one or more sets of the overclocked memory settings. The one or more sets of the overclocked memory settings are tested for stability of the memory and a stability indication is output for each of the one or more sets of the overclocked memory settings. One of the one or more sets of the overclocked memory settings are selected as optimized overclocked memory settings for the memory.

Abstract: In accordance with described techniques for processing-in-memory (PIM) search stop control, a computing system or computing device includes a memory system that includes a stop condition check component, which receives an instruction that includes a programmed check value. The stop condition check component compares the programmed check value to outputs of a PIM component, and the stop condition check component initiates a stop instruction to stop the PIM component from processing subsequent computations based on an output of the PIM component matching the programmed check value.

Abstract: Induced signal marginality for random number generation is described. In accordance with the described techniques, a pseudorandom number is transmitted across an interface while the interface is operated with settings configured to cause instability in the interface. A random number is received as an output of the interface. The settings configured to cause instability in the interface include overclocked settings of interface operating parameters.

Abstract: In accordance with described techniques for PE-centric all-to-all communication, a distributed computing system includes processing elements, such as graphics processing units, distributed in clusters. An all-to-all communication procedure is performed by the processing elements that are each configured to generate data packets in parallel for all-to-all data communication between the clusters. The all-to-all communication procedure includes a first stage of intra-cluster parallel data communication between respective processing elements of each of the clusters; a second stage of inter-cluster data exchange for all-to-all data communication between the clusters; and a third stage of intra-cluster data distribution to the respective processing elements of each of the clusters.

Abstract: A technique for performing ray tracing operations is provided. The technique includes determining error bounds for a rotation operation for a ray; selecting a technique for determining whether the ray intersects a bounding box based on the error bounds; and determining whether the ray hits the bounding box based on the selected technique.

Abstract: Resource access control is described. In accordance with the described techniques, a process (e.g., an application process, a system process, etc.) issues an instruction seeking access to a computation resource (e.g., a processor resource, a memory resource, etc.) to perform a computation task. An execution context for the instruction is checked to determine whether the execution context includes a resource indicator indicating permission to access the processor resource. Alternatively or additionally, the instruction is checked against an access table which identifies processes that are permitted and/or not permitted to access the computation resource.

Abstract: A page swapping memory protection system tracks accesses to physical memory pages, such as in a table with each row storing a physical memory page address and a counter value. This counter value records the number of accesses (e.g., read access or write accesses) to the corresponding physical memory page. In response to one of the counters exceeding a threshold value, the corresponding physical memory page is swapped with another page in physical memory (e.g., a page the table indicates has a smallest number of accesses). According, unreliability in the physical memory introduced due to frequent accesses to a particular physical memory page is mitigated.

Abstract: Systems and methods for building applications by automatically incorporating application performance data into the application build process are disclosed. By capturing build settings and performance data from prior applications being executed on different computing systems such as bare metal and virtualized cloud instances, a performance database may be maintained and used to predict build settings that improve application performance (e.g., on a specific computing system or computing system configuration).

Abstract: A processing system including a parallel processing unit selectively allocating pages of memory for interleaving across configurable subsets of channels based on a mode of allocation. In some embodiments, in a first mode, a page of memory is allocated to and interleaved across a plurality of channels, and in a second mode, a page of memory is allocated to and interleaved across a subset of the plurality of channels.

Abstract: A cache stores, along with data that is being transferred from a higher level cache to a lower level cache, information indicating the higher level cache location from which the data was transferred. Upon receiving a request for data that is stored at the location in the higher level cache, a cache controller stores the higher level cache location information in a status tag of the data. The cache controller then transfers the data with the status tag indicating the higher level cache location to a lower level cache. When the data is subsequently updated or evicted from the lower level cache, the cache controller reads the status tag location information and transfers the data back to the location in the higher level cache from which it was originally transferred.

Abstract: A driver circuit includes a feed-forward equalization (FFE) circuit. The FFE circuit receives a plurality of pulse-amplitude modulation (PAM) symbol values to be transmitted at one of multiple PAM levels. The FFE circuit includes a first partial lookup table, one or more additional partial lookup tables, and an adder circuit. The first partial lookup table contains partial finite impulse-response (FIR) values and indexed based on a current PAM symbol value, a precursor PAM symbol value, and a postcursor PAM symbol value. The one or more additional partial lookup tables each contain partial FIR values and indexed based on a respective additional one or more of the PAM symbol values. The adder circuit adds results of lookups from the first partial lookup table and the additional partial lookup tables to produce an output value.

Abstract: In accordance with described techniques for DRAM row management for processing in memory, a plurality of instructions are obtained for execution by a processing in memory component embedded in a dynamic random access memory. An instruction is identified that last accesses a row of the dynamic random access memory, and a subsequent instruction is identified that first accesses an additional row of the dynamic random access memory. A first command is issued to close the row and a second command is issued to open the additional row after the row is last accessed by the instruction.

Abstract: A memory request issue counter (MRIC) is maintained that is incremented for every memory access a central processing unit core makes. A region reuse distance table is also maintained that includes multiple entries each of which stores the region reuse distance for a corresponding region. When a memory access request for a physical address is received, a reuse distance for the physical address is calculated. This reuse distance is the difference between the current MRIC value and a previous MRIC value for the physical address. The previous MRIC value for the physical address is the MRIC value the MRIC had when a memory access request for the physical address was last received. A region reuse distance for a region that includes the physical address is generated based on the reuse distance for the physical address and used to manage the cache.

Abstract: A distributed cache network used for machine learning is provided which comprises a network fabric having file systems which store data and a plurality of processing devices, each comprising cache memory and a processor configured to execute a training of a machine learning model and selectively cache portions of the data based on a frequency with which the data is accessed by the processor. Each processing device stores metadata identifying portions of data which are cached in the cache memory and other portions of the data which are cached in other processing devices of the network. When requested data is not cached in another processing device, the portion of requested data is accessed from a network file system via a client to server channel and is accessed from another processing device via a client to client channel when the requested data is cached in the other processing device.

Abstract: In accordance with the described techniques for condensed coherence directory entries for processing in memory, a computing device includes a core that includes a cache, a memory that includes multiple banks, a coherence directory that includes a condensed entry indicating that data associated with a memory address and the multiple banks is not stored in the cache, and a cache coherence controller. The cache coherence controller receives a processing-in-memory command to the memory address and performs a single lookup in the coherence directory for the processing-in-memory command based on inclusion of the condensed entry in the coherence directory.