Abstract: Physical therapy assistant-as-a-service (PTaaS) enables the automatic evaluation of a patient's performance of physical therapy exercises and the automatic provision of feedback to the patient on their exercise performance in real-time. A patient device can provide real-time patient exercise video to a PTaaS backend that performs checks prior to the patient performing the exercise (pre-checks) and checks during patient performance of the exercise (live checks). If any of the checks fail, the PTaaS can provide feedback to the patient, such as if the patient is in an incorrect starting pose or has a body part at an incorrect angle before beginning the exercise or if the patient's form or posture during performance of the exercise needs to be adjusted. The PTaaS can automatically generate exercise metrics, reports, and physical therapy insights that a physical therapy clinician can access from a clinician portal.

Abstract: Systems, apparatuses, and methods include technology that determines, with a neural network, that a first eviction node stored in a cache will be evicted from the cache based on a cache policy. The first eviction node is part of a plurality of nodes associated with a graph. Further, a subset of nodes of the plurality of nodes remains in the cache after the eviction of the first eviction node from the cache. The technology further tracks a number of cache hits on the cache during an aggregation operation associated with a hardware accelerator, where the aggregation operation is executed on the subset of nodes that remain in the cache after the eviction of the eviction node from the cache. The technology executes a training process on the neural network to adjust the cache policy based on the number of the cache hits.

Abstract: Apparatus, devices, systems, methods, and articles of manufacture are disclosed to allocate power in a computing device. An example system includes a compiler to: analyze power consumption behavior of power consumption units of the computing device; build a power profile; and generate source code with hints of the power profile. The example system includes a power control circuit to: develop a power policy using the hints of the power profile and requests for power licenses from the power consumption units of the computing device; and allocate power to the power consumption units based on the power profile.

Abstract: Methods and apparatus relating to autonomous C state mechanism and computational engine alignment for improved processor power efficiency. are described. An embodiment determines whether a semiconductor package should enter a package C state based on energy consumption values for entry into and exit from the package C state, an amount of time the semiconductor package stayed in the package C state previously, and one or more breakeven time points between the various package C states. Another embodiment detects a delay by an imaging computational unit of a processor to enter a low power consumption state relative to one or more other computational units of the processor. The logic causes the imaging computational unit to enter the low power consumption state in response to detection of the delay. Other embodiments are also disclosed and claimed.

Abstract: A medical waste treatment system, comprising: (1) A main treatment unit comprising: a waste receiver cover comprising a UV lamp; a waste shredding unit comprising a shredding bin and at least two rotating shredding blades; a compression unit to push the medical waste towards the shredding blades; a separator to allow the passage of shredded medical waste of a predefined shape and diameter; and a disinfectant delivery unit comprising a plurality of nozzles. (2) A disinfection unit which comprises a disinfectant mixing bin and a mixing unit. (3) A main liquid management unit comprising a disinfectant delivery unit that comprises a plurality of nozzles interconnected to the main treatment unit. (4) A separator unit to separate the liquid from shredded waste; and a centrifuge to further remove liquid from the shredded waste.

Abstract: Apparatus, devices, systems, methods, and articles of manufacture are disclosed to allocate power in a computing device. An example system includes a compiler to: analyze power consumption behavior of power consumption units of the computing device; build a power profile; and generate source code with hints of the power profile. The example system includes a power control circuit to: develop a power policy using the hints of the power profile and requests for power licenses from the power consumption units of the computing device; and allocate power to the power consumption units based on the power profile.

Abstract: Methods and apparatus relating to autonomous C state mechanism and computational engine alignment for improved processor power efficiency. are described. An embodiment determines whether a semiconductor package should enter a package C state based on energy consumption values for entry into and exit from the package C state, an amount of time the semiconductor package stayed in the package C state previously, and one or more breakeven time points between the various package C states. Another embodiment detects a delay by an imaging computational unit of a processor to enter a low power consumption state relative to one or more other computational units of the processor. The logic causes the imaging computational unit to enter the low power consumption state in response to detection of the delay. Other embodiments are also disclosed and claimed.

Abstract: Techniques related to multi-block memory reads for image de-warping are discussed. Such techniques may include copying pixel data corresponding to overlapping regions of memory between adjacent image modification regions from a local buffer and retrieving pixel data corresponding to non-overlapping regions from external memory.

Abstract: Techniques related to multi-block memory reads for image de-warping are discussed. Such techniques may include copying pixel data corresponding to overlapping regions of memory between adjacent image modification regions from a local buffer and retrieving pixel data corresponding to non-overlapping regions from external memory.

Abstract: Methods and apparatus relating to autonomous C state mechanism and computational engine alignment for improved processor power efficiency. are described. An embodiment determines whether a semiconductor package should enter a package C state based on energy consumption values for entry into and exit from the package C state, an amount of time the semiconductor package stayed in the package C state previously, and one or more breakeven time points between the various package C states. Another embodiment detects a delay by an imaging computational unit of a processor to enter a low power consumption state relative to one or more other computational units of the processor. The logic causes the imaging computational unit to enter the low power consumption state in response to detection of the delay. Other embodiments are also disclosed and claimed.

Abstract: Techniques related to multi-block memory reads for image de-warping are discussed. Such techniques may include copying pixel data corresponding to overlapping regions of memory between adjacent image modification regions from a local buffer and retrieving pixel data corresponding to non-overlapping regions from external memory.

Abstract: Techniques related to multi-block memory reads for image de-warping are discussed. Such techniques may include copying pixel data corresponding to overlapping regions of memory between adjacent image modification regions from a local buffer and retrieving pixel data corresponding to non-overlapping regions from external memory.

Abstract: A processor device processes data samples of a radio signal in a mobile communication system. A fast flow process is executed for all samples and a batch process is executed at intervals on a subset of the samples. The device has a processor for executing the flow process via a local buffer memory, a memory interface to a system memory, and a memory controller for controlling storing of the data samples in the buffer memory. The processor establishes whether data samples in the local buffer memory are part of the subset, and if not, invalidates them after executing the flow process. The memory controller provides free memory space in the local buffer by transferring data samples which are not invalidated from the local buffer memory to the system memory, and by invalidating processed samples. Advantageously the local buffer may be relatively small, while the amount of data transferred to the system memory is limited.

Abstract: A Common Public Radio Interface, CPRI, lane controller of a processor, in a Time Division Duplex, TDD, system, said CPRI lane controller comprising: a Direct Memory Access (or more than one), DMA, controller connected to a memory through a switch fabric to perform read or/and write memory access transactions via an internal system bus of said processor, wherein said DMA controller is adapted to generate a RX/TX transaction interrupt(s) for each completed memory access RX/TX transaction counted by a corresponding transaction counter(s) which provides a TDD slot awareness interrupt(s) when a RX/TX TDD slot has terminated, wherein said DMA controller has a steering control(s) adapted to steer the memory access transactions either to said memory or to be legitimately blocked by said switch fabric in response to said TDD slot awareness interrupt(s) to save bandwidth, BW, of the internal system bus of said processor.

Abstract: A processor device processes data samples of a radio signal in a mobile communication system. A fast flow process is executed for all samples and a batch process is executed at intervals on a subset of the samples. The device has a processor for executing the flow process via a local buffer memory, a memory interface to a system memory, and a memory controller for controlling storing of the data samples in the buffer memory. The processor establishes whether data samples in the local buffer memory are part of the subset, and if not, invalidates them after executing the flow process. The memory controller provides free memory space in the local buffer by transferring data samples which are not invalidated from the local buffer memory to the system memory, and by invalidating processed samples. Advantageously the local buffer may be relatively small, while the amount of data transferred to the system memory is limited.

Abstract: A system for synchronizing and re-ordering data transmitted between first and second clock domains associated with first and second device interfaces, respectively, includes a splitter, an arbiter, a transaction manager, and a read data buffer. The splitter receives a parent read request from one or more data input ports of the first device interface and splits it into one or more read requests. The arbiter receives the one or more read requests and selects one of the read requests and transmits it to the transaction manager. The transaction manager allocates an entry to the read request and then the read request is transmitted to the read data buffer. Thereafter, the read data buffer transmits the read request to the second device interface and transmits received response data to the first device interface.

Abstract: Methods and apparatus relating to autonomous C state mechanism and computational engine alignment for improved processor power efficiency are described. An embodiment determines whether a semiconductor package should enter a package C state based on energy consumption values for entry into and exit from the package C state, an amount of time the semiconductor package stayed in the package C state previously, and one or more breakeven time points between the various package C states. Another embodiment detects a delay by an imaging computational unit of a processor to enter a low power consumption state relative to one or more other computational units of the processor. The logic causes the imaging computational unit to enter the low power consumption state in response to detection of the delay. Other embodiments are also disclosed and claimed.

Abstract: A Common Public Radio Interface, CPRI, lane controller of a processor, in a Time Division Duplex, TDD, system, said CPRI lane controller comprising: a Direct Memory Access (or more than one), DMA, controller connected to a memory through a switch fabric to perform read or/and write memory access transactions via an internal system bus of said processor, wherein said DMA controller is adapted to generate a RX/TX transaction interrupt(s) for each completed memory access RX/TX transaction counted by a corresponding transaction counter(s) which provides a TDD slot awareness interrupt(s) when a RX/TX TDD slot has terminated, wherein said DMA controller has a steering control(s) adapted to steer the memory access transactions either to said memory or to be legitimately blocked by said switch fabric in response to said TDD slot awareness interrupt(s) to save bandwidth, BW, of the internal system bus of said processor.

Abstract: A system for synchronizing and re-ordering data transmitted between first and second clock domains associated with first and second device interfaces, respectively, includes a splitter, an arbiter, a transaction manager, and a read data buffer. The splitter receives a parent read request from one or more data input ports of the first device interface and splits it into one or more read requests. The arbiter receives the one or more read requests and selects one of the read requests and transmits it to the transaction manager. The transaction manager allocates an entry to the read request and then the read request is transmitted to the read data buffer. Thereafter, the read data buffer transmits the read request to the second device interface and transmits received response data to the first device interface.

Abstract: A method for selecting a cache way, the method includes: selecting an initially selected cache way out of multiple cache ways of a cache module for receiving a data unit; the method being characterized by including: searching, if the initially selected cache way is locked, for an unlocked cache way, out of at least one group of cache ways that are located at predefined offsets from the first cache way.