Nvidia Corporation

Carrier Corporation

Yilin Zhang

Yan Zhou

Weijuan Wang

Qifei Fan

Jin Liu

Hao Fang Wu

Ujval J Kapasi

Xin Hao Cheng

Yuling Shi

Xin Cheng

HongYi Li

Dilip Sequeira

Alex Fit-Florea

Prashant Gaikwad

Slawomir Kierat

Geng Chen

Dongsheng Yang

Shangang Zhang

Zacky Zhou

Szymon Migacz

Maxim Milakov

A heat pump system comprises: a compressor having a compressor inlet and a compressor outlet; a change-over valve configured to selectively connect the compressor inlet and the compressor outlet to a first flow path and a second flow path; a heat-source-side heat exchanger on the first flow path; a user-side heat exchanger on the second flow path; a first branch and a second branch between the first flow path and the second flow path, the first branch having a three-way valve, and the second branch having a two throttling devices, wherein the three-way valve has two ports for communicating with the first flow path and a third port for communicating with the second flow path; and a third branch connected between the first throttling device and the second throttling device and the second port of the three-way valve, where an economizer is provided on the third branch.

Heat pump system and control method thereof

Apparatuses, systems, and techniques for scheduling deep learning tasks in hardware are described. One accelerator circuit includes multiple fixed-function circuits that each processes a different layer type of a neural network. A scheduler circuit receives state information associated with a respective layer being processed by a respective fixed-function circuit and dependency information that indicates a layer dependency condition for the respective layer. The scheduler circuit determines that the layer dependency condition is satisfied using the state information and the dependency information and enables the fixed-function circuit to process the current layer at the respective fixed-function circuit.

Hardware circuit for deep learning task scheduling

Apparatuses, systems, and techniques to detect faults in processing pipelines are described. One accelerator circuit includes a fixed-function circuit that performs an operation corresponding to a layer of a neural network. The fixed-function circuit includes a set of homogeneous processing units and a fault scanner circuit. The fault scanner circuit includes an additional homogeneous processing unit to scan each processing unit of the set for functional faults in a sequence.

Hardware-based fault scanner to detect faults in homogeneous processing units

An oil separator and an air conditioning system provided with the oil separator are disclosed by the present disclosure. The oil separator comprises: an inlet () configured to receive a refrigerant; a duct () through which the refrigerant flows, the duct having a circumferential wall and comprising: a first section comprising a first muffler adjacent to the inlet, the first muffler () comprising a chamber () defined by a first wall () of the first section and designed to have a length dimension (L) in a length direction of the duct and a depth dimension (H) intersecting the length dimension; and a second section downstream of the first section, a second wall () of the second section being provided with a second muffler (); and an oil separation assembly () through which the refrigerant passing through the duct passes. The present disclosure has a simple and reliable structure and is easy to implement.

Oil separator with integrated muffler

A heat pump system and a control method thereof. The heat pump system includes: a major heat exchange loop (), including at least one compressor (), a flow-path switching valve (), a condenser (), a first throttling element (), an economizer (), and an evaporator () that are connected sequentially to form a loop; and an air supply branch (), which is connected from a flow path between the first throttling element and the economizer to an air supply inlet of the compressor, the air supply branch being provided with a switch valve () for preventing a gas-phase refrigerant from flowing back; where a pressure balance branch () is further included, which is connected from the air supply branch at the upstream of the switch valve to a low-pressure gas-phase refrigerant side of the major heat exchange loop.

One method to mitigate vibration and sound level in heat pump chiller with evi function

Aspects of the present invention are directed to computer-implemented techniques for performing data compression and conversion between data formats of varying degrees of precision, and more particularly for improving the inferencing (application) of artificial neural networks using a reduced precision (e.g., INT8) data format. Embodiments of the present invention generate candidate conversions of data output, then employ a relative measure of quality to identify the candidate conversion with the greatest accuracy (i.e., least divergence from the original higher precision values). The representation can be then be used during inference to perform computations on the resulting output data.

Automated methods for conversions to a lower precision data format

Growing community of inventors

Shanghai, China

Yilin Zhang