Abstract: Techniques described herein address the above challenges that arise when using host executed software to manage vector databases by providing a vector database accelerator and shard management offload logic that is implemented within hardware and by software executed on device processors and programmable data planes of a programmable network interface device. In one embodiment, a programmable network interface device includes infrastructure management circuitry configured to facilitate data access for a neural network inference engine having a distributed data model via dynamic management of a node associated with the neural network inference engine, the node including a database shard of a vector database.

Abstract: System and techniques for power-based adaptive hardware reliability on a device are described herein. A hardware platform is divided into multiple partitions. Here, each partition includes a hardware component with an adjustable reliability feature. The several partitions are placed into one of multiple reliability categories. A workload with a reliability requirement is obtained and executed on a partition in a reliability category that satisfies the reliability requirements. A change in operating parameters for the device is detected and the adjustable reliability feature for the partition is modified based on the change in the operating parameters of the device.

Abstract: An apparatus includes a host interface, a network interface, and programmable circuitry communicably coupled to the host interface and the network interface, the programmable circuitry comprising one or more processors are to implement network interface functionality and are to receive a prompt directed to an artificial intelligence (AI) model hosted by a host device communicably coupled to the host interface, apply a prompt tuning model to the prompt to generate an initial augmented prompt, compare the initial augmented prompt for a match with stored data of a prompt augmentation tracking table comprising real-time datacenter trend data and cross-network historical augmentation data from programmable network interface devices in a datacenter hosting the apparatus, generate, in response to identification of the match with the stored data, a final augmented prompt based on the match, and transmit the final augmented prompt to the AI model.

Abstract: Various embodiments described herein relate to systems and methods for assessing a condition of a package in a facility. In this regard, first data from one or more first sensors is received once a worker picks the package at a first location in the facility. Further, based on the first data it is determined if the worker is holding the package. If the worker holds the package, second data is received from one or more second sensors such that the one or more second sensors are different from the one or more first sensors. Based on the second data, a package handling pattern for the package is determined. One or more notifications are provided to a mobile device associated with the worker if it is determined that the package handling pattern causes damage to the package.

Abstract: A method for providing one or more customized experience to a user profile associated with an application is disclosed. The method comprises querying a key-value pair store for retrieving a key-value pair associated with the user profile. The retrieved key-value pair is loaded into a first memory. Further, one or more segment definitions for the one or more segments are received from a second memory. The one or more segment definitions are based on at least one of the user profile, user behaviour, user transaction on the application, user interaction with the application, and user subscription. Further, the one or more segment definitions and the one or more user events are evaluated in the first memory. The customized experience is provided to the user profile within sub-second latency and the customized experience is updated based on change in the one or more user events associated with the user profile.

Abstract: A disclosed method reduces memory consumption of a trained sequential model. The method includes receiving, from a client application, an initial processing request identifying an input sequence to be processed by the trained sequential model and an initial value for an output size parameter specifying a requested size of output from the trained sequential model. The method further includes sequentially transmitting, to the trained sequential model, multiple partial processing requests based on the initial processing request that each specify a fraction of the initial value as the output size parameter and receiving a sequence of output responses from the trained sequential model generated in response to processing the multiple partial processing requests. The method further provides for returning, to the client application, a final merged response that includes the sequence of output responses.

Abstract: A method includes obtaining an indication of a user interface (UI) component of a user interface, and determining an association between the UI component and a dynamic identifier. The method also includes, based on determining the association, determining one or more static properties of one or more parent UI components of the UI component, and generating a component selector for the UI component based on the one or more static properties. The method further includes outputting the component selector for the UI component.

Abstract: A disclosed method facilitates an increase in utilization with respect to a resource quota allocated to a tenant from a shared resource pool. The method includes transmitting a lease request to a quota service on behalf of the tenant, where the lease request identifies a processing task and specifies quantity of cloud-based resources requested from the shared resource pool for execution of the processing task. The method further provides for determining, based on a feedback signal received from the quota service, whether grant of the lease request would cause the tenant to exceed a resource quota allocated to the tenant and dynamically decreasing parallelism of active tasks being processed by the cloud-based resources on behalf of the tenant in response to determining that grant of the lease request would cause the tenant to exceed the resource quota limit.

Abstract: A system provides capacity-based load balancing across model endpoints of a cloud-based artificial intelligence (AI) model. The system includes a consumption determination engine executable to determine a net resource consumption for processing tasks in a workload generated by a client application for input to the trained machine learning model. The system also includes a load balancer that determines a distribution of available resource capacity in a shared resource pool comprising compute resources at each of the multiple model endpoints. The load balancer allocates parallelizable tasks of the workload among the compute resources at the multiple model endpoints based on the net resource consumption of the tasks and on the distribution of available resource capacity in the shared resource pool.

Abstract: An architecture to perform resource management among multiple network nodes and associated resources is disclosed. Example resource management techniques include those relating to: proactive reservation of edge computing resources; deadline-driven resource allocation; speculative edge QOS pre-allocation; and automatic QoS migration across edge computing nodes.

Abstract: Methods and apparatus for software-defined coherent caching of pooled memory. The pooled memory is implemented in an environment having a disaggregated architecture where compute resources such as compute platforms are connected to disaggregated memory via a network or fabric. Software-defined caching policies are implemented in hardware in a processor SoC or discrete device such as a Network Interface Controller (NIC) by programming logic in an FPGA or accelerator on the SoC or discrete device. The programmed logic is configured to implement software-defined caching policies in hardware for effecting disaggregated memory (DM) caching in an associated DM cache of at least a portion of an address space allocated for the software application in the disaggregated memory.

Abstract: Embodiments described herein are generally directed to the use of sidecars to perform dynamic Application Programming Interface (API) contract generation and conversion. In an example, a first sidecar of a source microservice intercepts a first call to a first API exposed by a destination microservice. The first call makes use of a first API technology specified by a first contract and is originated by the source microservice. An API technology is selected from multiple API technologies. The selected API technology is determined to be different than the first API technology. Based on the first contract, a second contract is dynamically generated that specifies an intermediate API that makes use of the selected API technology. A second sidecar of the destination microservice is caused to generate the intermediate API and connect the intermediate API to the first API.

Abstract: Techniques for expanded trusted domains are disclosed. In the illustrative embodiment, a trusted domain can be established that includes hardware components from a processor as well as an off-load device. The off-load device may provide compute resources for the trusted domain. The trusted domain can be expanded and contracted on-demand, allowing for a flexible approach to creating and using trusted domains.

Abstract: In an embodiment, network device apparatus is provided that includes packet processing circuitry to determine if target data associated with a memory access request is stored in a different device than that identified in the memory access request, and based on the target data associated with the memory access request identified as stored in a different device than that identified in the memory access request, may cause transmission of the memory access request to the different device. The memory access request may comprise an identifier of a requester of the memory access request and the identifier may comprise a Process Address Space identifier (PASID).

Abstract: Deployment of resources utilizing improved mixture of experts processing is described. An example of an apparatus includes one or more network ports; one or more direct memory access (DMA) engines; and circuitry for mixture of experts (MoE) processing in the network, wherein the circuitry includes at least circuitry to track routing of tokens in MoE processing, prediction circuitry to generate predictions regarding MoE processing, including predicting future token loads for MoE processing, and routing management circuitry to manage the routing of the tokens in MoE processing based at least in part on the predictions regarding the MoE processing.

Abstract: An apparatus for aircraft propulsion includes a propeller and a hybrid-cooled electric engine mounted to a support structure and configured to rotate the propeller. The electric engine is located in an enclosure, with a first heat transfer element thermally coupled to the electric engine via a fluid flow path, partially located outside the enclosure. A second heat transfer element, integral to or thermally coupled to the electric engine, provides air cooling. Both heat transfer elements are housed within the aircraft's support structure. At least one air inlet on the upper side of the support structure receives propeller downwash. A first cooling path directs a portion of the downwash from the air inlet to the first heat transfer element, and a first air outlet exhausts the downwash.

Abstract: An apparatus includes a host interface, a network interface, and a programmable circuitry communicably coupled to the host interface and the network interface. The programmable circuitry can include one or more processors to implement network interface functionality, and a discrete trusted platform module (dTPM) to enable the one or more processors to establish a secure boot mechanism for the apparatus, wherein the one or more processors are to instantiate a virtual TPM (vTPM) manager that is associated with the dTPM, the vTPM manager to host vTPM instances corresponding to one or more virtualized environments hosted on at least one of the programmable circuitry or a host device communicable coupled to the apparatus.

Abstract: A method includes receiving, via a graphical user interface (GUI), by a processing device, a first user input to view data associated with a first process chamber in a first chamber data mode. Data of the first chamber data mode includes data of a process operation performed in the first process chamber. The method further includes providing, for display on the GUI, first display data of the first data chamber mode responsive to receiving the first user input. The method further includes receiving a second user input to view data associated with the first process chamber in a second chamber data mode. Data of the second chamber data mode includes data of a virtual process operation performed by a virtual representation of the first process chamber. The method further includes providing, for display on the GUI, second display data of the second data chamber mode.

Abstract: A cooling arrangement for conditioning air within a cabin of an underground work machine includes blowers to generate an air flow, a first evaporator, and a second evaporator. The first evaporator provides passage to a first portion of the air flow and receive a first coolant to facilitate a first heat transfer from the first portion of the air flow to the first coolant. The second evaporator provides passage to a second portion of the air flow and receive a second coolant to facilitate a second heat transfer from the second portion of the air flow to the second coolant. A receipt of the first coolant into the first evaporator for the first transfer and a receipt of the second coolant into the second evaporator for the second transfer occurs either independently or simultaneously with each other for supplying conditioned air flow into the cabin to cool the cabin.

Abstract: A VTOL aircraft includes a plurality of lift propellers configured to rotated by lift motors to provide vertical thrust during takeoff, landing and hovering operations. The lift propellers are configured to generate a cooling airflow to cool the lift motors during use. During a cruise operation when the VTOL aircraft is in forward motion, the lift propellers may be stowed in a stationary position. Therefore, the cooling airflow may be reduced or eliminated when it is not needed.