Abstract: Technologies for enabling remote direct memory access (RDMA) transport of serialized objects in streaming pipelines are described. One method of a first computing device that stores a serialized object in a first memory can generate a remote descriptor associated with the serialized object. The remote descriptor uniquely identifies the location of the serialized object and a reference count token. The first computing device sends the remote descriptor to a second computing device in the data center over a network fabric. The second computing device uses the remote descriptor to obtain the contiguous block from the first memory for storage at a second memory associated with the second computing device. The value of the reference count token can be updated by receiving a message from the second computing device, and the remote descriptor can be released responsive to the value of the reference count token satisfying a threshold value.

Abstract: A storage system (400) includes a storage processor (420-1) such as a storage card resident in a host server (410-1) and coupled to the storage device (450-1). The storage processor (420-1) may be configured to create a virtual volume (428-1), store content derived from an image (492) downloaded from a URL storage corresponding to the virtual volume (428-1), and present the virtual volume (428-1) to the host server as a boot LUN. A management infrastructure (480) can be used to create a library (490) of images (492, 494) corresponding to different storage system characteristics and used to selected which URL is provided to the storage processor (420-1).

Abstract: In various examples, a deep neural network (DNN) is trained to accurately predict, in deployment, distances to objects and obstacles using image data alone. The DNN may be trained with ground truth data that is generated and encoded using sensor data from any number of depth predicting sensors, such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. Camera adaptation algorithms may be used in various embodiments to adapt the DNN for use with image data generated by cameras with varying parameters—such as varying fields of view. In some examples, a post-processing safety bounds operation may be executed on the predictions of the DNN to ensure that the predictions fall within a safety-permissible range.

Abstract: In various examples, a VPU and associated components may be optimized to improve VPU performance and throughput. For example, the VPU may include a min/max collector, automatic store predication functionality, a SIMD data path organization that allows for inter-lane sharing, a transposed load/store with stride parameter functionality, a load with permute and zero insertion functionality, hardware, logic, and memory layout functionality to allow for two point and two by two point lookups, and per memory bank load caching capabilities. In addition, decoupled accelerators may be used to offload VPU processing tasks to increase throughput and performance, and a hardware sequencer may be included in a DMA system to reduce programming complexity of the VPU and the DMA system. The DMA and VPU may execute a VPU configuration mode that allows the VPU and DMA to operate without a processing controller for performing dynamic region based data movement operations.

Abstract: Technologies for batching remote descriptors of serialized objects in streaming pipelines are described. One method of a first computing device generates a streaming batch of remote descriptors. Each remote descriptor uniquely identifies a contiguous block of a serialized object. The first computing device sends at least one of the remote descriptors to a second computing device before the streaming batch is completed. At least some contents of a contiguous block are obtained for storage at a second memory associated with the second computing device before the streaming batch is completed.

Abstract: Embodiments relate to methods for efficiently encoding sensor data captured by an autonomous vehicle and building a high definition map using the encoded sensor data. The sensor data can be LiDAR data which is expressed as multiple image representations. Image representations that include important LiDAR data undergo a lossless compression while image representations that include LiDAR data that is more error-tolerant undergo a lossy compression. Therefore, the compressed sensor data can be transmitted to an online system for building a high definition map. When building a high definition map, entities, such as road signs and road lines, are constructed such that when encoded and compressed, the high definition map consumes less storage space. The positions of entities are expressed in relation to a reference centerline in the high definition map. Therefore, each position of an entity can be expressed in fewer numerical digits in comparison to conventional methods.

Abstract: The present disclosure provides methods, systems, and computer program products that use deep learning models to classify candidate mutations detected in sequencing data, particularly suboptimal sequencing data. The methods, systems, and programs provide for increased efficiency, accuracy, and speed in identifying mutations from a wide range of sequencing data.

Abstract: Systems and methods in accordance with the present disclosure can implement a parallel processing system, such as a graphics processing unit (GPU)-based system, to generate solutions to complex computational problems. Aspects of this technical solution can retrieve a plurality of solutions each representing a plurality of values in a multi-dimensional space, allocate, to one or more processing units associated with the one or more circuits and having a parallelized configuration, one or more of the plurality of solutions, modify, by the one or more processing units according to the parallelized configuration, at least one value of the one or more solutions allocated to the one or more processing units, to determine a plurality of modified solutions, and output, from the plurality of modified solutions, according to one or more criteria indicating a diversity among the plurality of modified solutions, a selected solution.

Abstract: In various examples, sets of testing data may be selected and applied to an MLM such that differences in performance of the MLM in the testing between the sets indicates and may be used to determine whether and/or an extent by which the MLM is trained to rely on artifacts. Training data for the MLM may be generated using a first value of a parameter that defines a value of a characteristic of the training data. For testing, first testing data may be selected that corresponds to a second value of the parameter that shifts the value in a first direction and second testing data may be selected that corresponds to a third value of the parameter that shifts the value in a second direction (e.g., opposite the first direction). Various possible actions may be taken based on results of analyzing the differences in performance.

Abstract: Systems and methods for cooling a datacenter are disclosed. In at least one embodiment, a server tray or box includes a surface with a flow-through fixture or manifold that extends on both sides of the surface, that includes an inward coupling, an outward coupling, a flow controller, and a state sensor, the state sensor to monitor a flow-through fixture or manifold, where a flow controller of a flow-through fixture or manifold can change a flow of a coolant through a flow-through fixture or manifold and can selectively trap a portion of a coolant within a flow-through fixture or manifold.

Abstract: Technologies for enabling downstream components to update upstream states in streaming pipelines are described. One method of a first computing device receives a remote promise object assigned to a first serialized object from a second computing device in the data center over a network fabric. The remote promise object uniquely identifies a first contiguous block of the first serialized object stored in a memory associated with the second computing device. The method obtains contents of the first contiguous block and sends contents of a second serialized object back to the second computing device to release the remote promise object.

Abstract: In various examples, sensor data—such as masked sensor data—may be used as input to a machine learning model to determine a confidence for object to person associations. The masked sensor data may focus the machine learning model on particular regions of the image that correspond to persons, objects, or some combination thereof. In some embodiments, coordinates corresponding to persons, objects, or combinations thereof, in addition to area ratios between various regions of the image corresponding to the persons, objects, or combinations thereof, may be used to further aid the machine learning model in focusing on important regions of the image for determining the object to person associations.

Abstract: Mechanisms to place flip-flops and other synchronous logic cells in a circuit layout in a clock on-chip variation-aware, predetermined order based on analysis of the clock gating, connectivity, and logic depth of the unplaced netlist. The resulting placements enable clock trees having a regular structure leading to improvements in clock on-chip variation, timing, and clock power.

Abstract: In various examples, systems and methods are disclosed for weighting one or more optional paths based on obstacle avoidance or other safety considerations. In some embodiments, the obstacle avoidance considerations may be computed using a comparison of trajectories representative of safety procedures at present and future projected time steps of an ego-vehicle and other actors to ensure that each actor is capable of implementing their respective safety procedure while avoiding collisions at any point along the trajectory. This comparison may include filtering out a path(s) of an actor at a time step(s)—e.g., using a one-dimensional lookup—based on spatial relationships between the actor and the ego-vehicle at the time step(s). Where a particular path—or point along the path—does not satisfy a collision-free standard, the path may be penalized more negatively with respect to the obstacle avoidance considerations, or may be removed from consideration as a potential path.

Abstract: Methods and structures are described for detecting clock anomalies, including anomalies in which the clock oscillates at a faster than expected rate or exhibits a shorter than expected clock phase instance. Example methods include starting a timer responsive to the start of a clock phase, wherein the timer duration is shorter than an expected duration of the clock phase. If the clock phase ends before the timer expires, a fast clock detection signal is asserted. Example structures include fast clock detection logic coupled to a clock signal. The logic includes a timer, circuitry to start the timer responsive to the clock signal entering a monitored phase, and error detection circuitry to assert a fast clock detection output if the monitored phase ends before the timer expires. In some embodiments, the timer duration may be based on a measured duration of a previous clock phase.

Abstract: Manufacturers perform tests on chips before the chips are shipped to customers. However, defects can occur on a chip after the manufacturer testing and when the chips are used in a system or device. The defects can occur due to aging or the environment in which the chip is employed and can be critical; especially when the chips are used in systems such as autonomous vehicles. To verify the structural integrity of the IC during the lifetime of the product, an in-system test (IST) is disclosed. The IST enables self-testing mechanisms for an IC in working systems. The IST mechanisms provide structural testing of the ICs when in a functional system and at a manufacturer's level of testing. Unlike ATE tests that are running on a separate environment, the IST provides the ability to go from a functional world view to a test mode.

Abstract: A performance metrics of a receiver is obtained using frames of an application hosted by a server that are received via a network. The one or more performance metrics include information indicative of a current occupancy of a frame buffer corresponding to the receiver and information indicative of a target occupancy of the frame buffer corresponding to the receiver. The frame buffer of the receiver is used to queue frames of the application for display. A frame rate associated with rendering at least one next frame of the application is adjusted using the one or more performance metrics of the receiver to control population of the frame buffer. Subsequent frames of the application hosted by the server are rendered using the adjusted frame rate. Upon rendering the subsequent frames, the server sends the subsequent frames to the receiver for display.

Abstract: Techniques for testing semiconductor devices include a semiconductor device having a plurality of components, a test bus, and a test data transfer unit. The test data transfer unit receives, from a host computer, configuration information for performing a test of the semiconductor device, reads, via a high-speed data transfer link, test data associated with the test from memory of the host computer using direct memory access, sends the test data to the plurality of components via the test bus, causes one or more operations to be performed on the semiconductor device to effect at least a portion of the test, and after the one or more operations have completed, retrieves test results of the at least a portion of the test from the test bus and stores, via the high-speed data transfer link, the test results in the memory of the host computer using direct memory access.

Abstract: Systems and methods for cooling a datacenter are disclosed. In at least one embodiment, a control unit within a rack has a pump or compressor unit to cause two-phase fluid to circulate through a cold plate associated with a computing device and to circulate through a heat exchanger associated with a rear door of a rack, so as to dissipate heat from a computing device through a heat exchanger by a control unit within a rack.

Abstract: Systems and methods herein address reference frame selection in video streaming applications using one or more processing units to identify a frame of a sequence of frames as a blurred frame based at least in part on a first variance of motion (VoM) of the frame being less than or equal to an adaptive threshold that is based in part on a moving average of variance of motion (MAoV) determined using one or more reference frames.