Abstract: A dry powder inhalation device has a means to aerosolize powdered medicament and introduce the aerosolized medicament to an airflow directed to a user outlet. The device has a primary deagglomeration structure in the form of a baffle through which the airflow passes downstream of the means to aerosolize the powdered medicament and a secondary deagglomeration structure which is a tortuous passage through which the airflow passes downstream of the primary deagglomeration structure. The deagglomeration structures cause disruption to the airflow in their vicinity whereby the particles of the aerosolized powdered medicament are deagglomerated due to the primary deagglomeration structure and the secondary deagglomeration structure. The insert defines a tortuous passage which has a trajectory with a centre line, whereby the centre line has two or more major bends, involving a first point of inflection between a pair of adjacent bends of the two or more major bends.

Abstract: Using multiple overlays with a data processing array includes loading an application in a data processing array. The data processing array includes a plurality of compute tiles each having a processor. The application specifies kernels executable by the processors and implements stream channels that convey data to the plurality of compute tiles. During runtime of the application, a plurality of overlays are sequentially implemented in the data processing array. Each overlay implements a different mode of data movement in the data processing array via the stream channels. For each overlay implemented, a workload is performed by moving data to the plurality of compute tiles based on the respective mode of data movement.

Abstract: An oven (1000) having a body (1002) that includes a base, a ceiling, and side walls extending between the base and the ceiling, the side walls at least partly surrounding a cooking cavity (1030), the oven (1000) including: an impeller assembly (1060) mounted to a first side wall of the oven body, the impeller assembly (1060) including: an impeller (1062) that is rotatable to direct air flow within the cooking cavity (1030); a plurality of air guides (2006) that at least partially surround the impeller (1062), the air guides (2006) each defining a channel that extends generally transversely from a central axis of the impeller (1062); and a pair of heating elements (2010) located on either side of the impeller (1062), whereby air that is directed from the impeller (1062) travels transversely along the channels of the air guides (2006), across the heating elements (2010), and into the cooking cavity (1030).

Abstract: Examples herein describe techniques for adapting a multiplier array (e.g., a systolic array implemented in a processing core) to perform different dot products. The processing core can include data selection logic that enables different configurations of the multiplier array in the core. For example, the data selection logic can enable different configurations of the multiplier array while using the same underlying hardware. That is, the multiplier array is fixed hardware but the data selection can transmit data into the matrix multiplier such that it is configured to perform different length dot products, perform more dot products in parallel, or change its output precision. In this manner, the same underlying hardware (i.e., the multiplier array) can be reconfigured for different dot products which can result in much more efficient use of the hardware.

Abstract: Using multiple overlays with a data processing array includes loading an application in a data processing array. The data processing array includes a plurality of compute tiles each having a processor. The application specifies kernels executable by the processors and implements stream channels that convey data to the plurality of compute tiles. During runtime of the application, a plurality of overlays are sequentially implemented in the data processing array. Each overlay implements a different mode of data movement in the data processing array via the stream channels. For each overlay implemented, a workload is performed by moving data to the plurality of compute tiles based on the respective mode of data movement.

Abstract: Examples herein describe techniques for reducing the amount of memory used during weight sparsity. When decompressing the weights, the uncompressed weight data typically has many zero values. By knowing the location of these zero values (e.g., their indices in a weight matrix), the processor core can prune some of the activations (e.g., logically reduce the size of the activation matrix) which improves the efficiency of the processor core. In embodiments herein, the processor core includes logic for identifying the indices of the non-zero value after decompressing the compressed weights. These indices can then be used to prune the activations to improve the efficiency of the processor core.

Abstract: An oven (10) having a body (15) that includes a base (20), a ceiling (25), and a wall (30) extending between the base (20) and the ceiling (25), the wall (30) at least partly surrounding a cooking cavity (35), the base (20) including a base cavity (50) and an opening (55) that extends between the base cavity (50) and the cooking cavity (35); the oven (10) further including a removable container (65) having a floor (81) and sidewalls (82) that define a chamber (85) to hold fluids therein, the base cavity (50) adapted to receive the removable container (65), the removable container (65) being movable relative to the base cavity (50) between a storage position in which the removable container (65) is retained in the base cavity (50); and an open position in which the removable container (65) is at least partially exposed relative to the base cavity (50), the oven (10) further including an interface assembly (150) providing a connection between a fluid outlet (97) of the removable container (65) and the cooking c