132 results about "Supercomputer" patented technology

The invention provides a time-space attention-driven soil humidity prediction method in a domestic supercomputing environment, and the method comprises the steps: constructing a time-space feature dynamic fusion soil humidity prediction network model based on attention guidance; the attention-guided spatio-temporal feature dynamic fusion soil humidity prediction network model comprises a convolutional long-short term memory network basic framework and a spatio-temporal attention mechanism module. The introduction of the ConvLSTM network effectively integrates the time sequence and space information of the soil humidity, the potential space-time dependence in the data is fully utilized, the space-time attention mechanism enables the model to adaptively pay attention to important time periods and space regions by dynamically adjusting the weight between time steps, and the accuracy of the model is improved. The limitation of fixed feature selection in a traditional method is avoided, so that the prediction precision is improved. The model can better capture space-time dynamic feature information, further identifies the relationship between the influence factor and the soil humidity, and especially shows unique advantages in complex space-time feature processing and modeling of long-time sequence data.

The invention relates to a management method and device for a distributed collection system facing the supercomputing Internet and a storage medium, and the method comprises the steps: obtaining the heartbeat information of all collection nodes in the distributed collection system from a first database, and obtaining the management authority according to the heartbeat information of all collection nodes; and when it is determined that the management authority is obtained, an instance list corresponding to the distributed acquisition system is obtained from the second database, and acquisition tasks on all the acquisition nodes in the distributed acquisition system are managed according to the heartbeat information of all the acquisition nodes and the instance list. According to the method, in the process of managing the acquisition tasks in the distributed acquisition system by using the method, a decentration effect can be realized by adopting a mode of mutually monitoring all the acquisition nodes, and the management efficiency and stability of the whole system are improved. And each acquisition node performs task management in a manner of obtaining the management authority, so that management conflicts can be avoided.

The invention relates to an ultrasonic large model sparse tensor optimization training and pushing acceleration method based on new generation supercomputing, and the method comprises the steps: carrying out the fine tuning of an ultrasonic large model, enabling the weight tensor of each expert model in the large model to generate a structured block sparse mode, and obtaining an ultrasonic sparse large model; constructing a sparse block index table and an expert index table, and associating to form a joint coding table; deploying an ultrasonic sparse large model to each parallel group of the new-generation heterogeneous supercomputing cluster; performing multi-round iterative training on the large sparse model; in each round of iterative training, dividing the sample ultrasonic data of the current batch into each parallel group; in each parallel group, determining a target expert model needing to be activated based on the feature data of each sample ultrasonic data, and positioning a target computing node based on the joint coding table; and activating the target expert model on the target computing node and loading the corresponding non-zero weight block so as to compute the feature data, and updating the model weight parameter based on an output result obtained by computation. The method can accelerate the processing efficiency of the ultrasonic large model.

Systems and methods for intelligent processing of requests in an artificial intelligence system, including receiving an incoming request at an AI broker, evaluating the incoming request to determine routing characteristics, selecting selected specialized language models from a plurality of specialized language models based on the routing characteristics, computing performance metrics for the one or more selected specialized language models, routing the incoming request to the selected specialized language models based on the performance metrics, and generating a final result by receiving and processing results from the selected specialized language models.

A method for assigning tasks to LLMs using h-including receiving documents relevant to embeddings, generating context-aware prompts responsive to at least one of a received prompt, derived prompts, and the received knowledge documents, each context-aware prompt corresponding to a specialized task of the one or more specialized tasks, transmitting each context-aware prompt to a respective h- that is configured to specialize in processing prompts having a specialty corresponding to the specialized task of the context-aware prompt, and receiving results from the h-LLMs.

A system for answering queries using one or more families of large language models (h-LLMs) including a processor, a communication device, one or more h-LLMs implemented as microservices in cloud container environments and being accessible via a cloud service API, and software that is executable to operate an input broker having a broker API operable to receive a user prompt from a user interface, generate a plurality of derived prompts, transmit the plurality of derived prompts to one or more h-LLMs via the cloud service API, operate an output broker operable to receive a plurality of h-LLM results, process the plurality of LLM results at the output broker to generate a result, and transmit the result to the user interface via the broker API.

A method for generating a merged large language model (LLM) from input data including receiving derived prompts generated from a user prompt and relevant contexts, receiving the relevant contexts related to the generation of the plurality of derived prompts, providing derived prompts and the relevant contexts to one or more merged LLMs, generating a plurality of results, and sending the plurality of results to an output broker.

An extension for a supercomputer rack, the extension including a reinforcement frame, defining a rectangular through-opening, and at least one fluidic connection support mounted to the reinforcement frame. The at least one fluidic connection support includes an upstream circuit, configured to convey a refrigerant liquid to a plurality of connectors configured to be each connected via a cold flexible manifold to a cooling circuit of a compute blade of a server, and a downstream circuit, configured to collect the heated refrigerant liquid having passed through a compute blade of a server via a hot flexible manifold and a plurality of collection connectors. The extension is configured to be positioned at a rear end of the rack.

A method for training large language models (LLMs) and using them for inference including a base LLM, creating connected models while training the base LLM by coupling multiple trained base a parallel, series, or hybrid architecture, creating specialized connected language models for different specialized processing tasks by supplementally training two or more connected models different respective training sets, supplementally training a subset of connected language models by selectively routing one or more training data inputs to one or more specialized connected language models of the plurality of specialized connected language models responsive to at least one of accuracy optimization and task specialization, providing test prompt sets to the subset, evaluating an accuracy metric of the outputs from the subset, and routing prompts or derived prompts to the subset.

A system and method for creating a merged large language model (h-LLM) using a bagging approach including receiving input data at a computer system, creating a plurality of data subsets from the input data, training a plurality of h-LLMs, each h-LLM of the plurality of h-LLMs being trained on a respective data subset of the plurality of data subsets, creating a merged h-LLM by merging the plurality of h-LLMs, and outputting the merged h-LLM.

The invention comprises (i) a compilation method for automatically converting a single-threaded software program into an application-specific supercomputer, and (ii) the supercomputer system structure generated as a result of applying this method. The compilation method comprises: (a) Converting an arbitrary code fragment from the application into customized hardware whose execution is functionally equivalent to the software execution of the code fragment; and (b) Generating interfaces on the hardware and software parts of the application, which (i) Perform a software-to-hardware program state transfer at the entries of the code fragment; (ii) Perform a hardware-to-software program state transfer at the exits of the code fragment; and (iii) Maintain memory coherence between the software and hardware memories. If the resulting hardware design is large, it is divided into partitions such that each partition can fit into a single chip. Then, a single union chip is created which can realize any of the partitions.

A zettascale supercomputer may be configured using an array of servers organized in computer pods comprising 4-60 servers per pod. Each server includes computer modules immersed in a tank in which cooling water is flowing. Copper bus bars deliver power to each pod in a range of 4-180 MW. Cooling water is delivered to each pod in a range of 200-24,000 gallons per minute. Supercomputers having an operating power in a range of 4 MW-10 GW are described.

The invention discloses a multi-source cooperative green electric power energy optimization scheduling system, and relates to the technical field of electric power system computing power scheduling, and the system comprises an interaction construction module which constructs a direct connection signal channel of a supercomputing CPU and an intelligent computing GPU; the impedance matching module monitors a signal transmission reflection coefficient in real time and obtains signal transmission delay to realize impedance compensation; the supercomputing power supply layer configuration module regulates and controls output voltage in real time through a voltage monitoring unit; the intelligent calculation power supply layer configuration module adapts to the instantaneous peak current demand of the intelligent calculation module; the supercomputing water-cooling heat dissipation execution module is used for adjusting the flow of a water-cooling pump after receiving the temperature signal; the intelligent calculation air cooling linkage control module is used for collecting the real-time temperature of the intelligent calculation module and controlling the rotating speed of a speed-adjustable fan according to a fan rotating speed adjusting formula so as to realize pulse type heating adaptive heat dissipation; and the comprehensive evaluation module constructs a comprehensive evaluation model and evaluates the system performance. The problems of high transmission delay, mismatched heat dissipation and unstable power supply are solved.

The present invention provides a method for predicting supercomputer job duration based on semantics and timing, which relates to the field of supercomputers and solves the problem of ignoring semantic information in the job path and timing information between jobs in supercomputer job runtime prediction. The method first obtains job log data, distinguishes different user types through data grouping, uses different data storage methods to store the user's job log data, and forms a user model training set through coarse-grained clustering. Subsequently, a prediction model for job runtime is constructed and model training is performed by improving the BERT architecture. During the training process, a timing prediction method is combined so that the prediction model predicts the timing of new jobs based on timing information. After the training is completed, the user submits the job path information of the new job, and the prediction model determines the job category of the new job and outputs the job runtime. The present invention improves the prediction accuracy of job runtime, facilitating subsequent backfill scheduling.

The invention provides an internal interconnection and intercommunication small supercomputing system, device and cluster. The system comprises a programmable processing unit, a plurality of high-speed serial interconnection ports, a high-speed data transmission module and a multi-level heterogeneous storage module, the high-speed serial interconnection port is used for connecting the computing unit; the high-speed data transmission module is used for sending the received to-be-processed task to the cache space of the programmable processing unit; the programmable processing unit is used for distributing to-be-processed tasks to the plurality of computing units for task processing based on a preset distribution strategy and receiving task processing results; and the multi-level heterogeneous storage module is used for storing the received task processing result. The programmable processing unit serves as an architecture center, various high-speed interfaces are managed and coordinated in a unified mode, an architecture and a management mechanism for interconnection and intercommunication of internal components of the small supercomputing system are achieved, and the data transmission efficiency is improved. The programmable processing unit distributes tasks to a plurality of computing units for parallel processing, so that the computing efficiency and the computing power utilization rate are improved.

The application relates to the technical field of supercomputer clusters, and discloses a method and device for allocating vnc resources in a supercomputer cluster, which comprises the following steps: a control node splices a job script to obtain a job script file, and compares the vnc resource data of a display node and the computing resource of a computing node with preset scheduling conditions; if the vnc resource data and the computing resource meet the preset scheduling conditions, the job script file is sent to the computing node; the computing node executes the job script file to start a vnc service; after the execution of the job script file is completed, the computing node acquires and executes a user application program; the control node calls a job state in the computing node, sends a vnc service stop request to the display node based on the job state, and controls the display node to stop the vnc service. The application realizes the reasonable allocation of the vnc resources in the supercomputer cluster, solves the node collapse caused by the unreasonable allocation of the vnc resources, and improves the graphic display effect of the application program.

The application discloses a supercomputer load balancing strategy determination method and device based on a structure grid, equipment and a storage medium, relates to the technical field of computational fluid dynamics, and comprises the following steps: setting a processor of a supercomputer as a computing node, constructing a structure grid with parameters such as the number of grid units and the number of adjacent surfaces based on a master-slave core heterogeneous architecture, establishing a load balancing model in combination with a hardware architecture, utilizing the model to segment the structure grid under constraint conditions such as node calculation time, inter-node communication, inter-core group communication and inter-slave core communication, obtaining segmented grid blocks, determining vertices and their weights according to the segmentation result, determining edges and their weights based on an adjacent relationship, setting the number of adjacent grid blocks as the vertex degree, thereby constructing a directed graph mapping result, processing the mapping result by using a graph partitioning algorithm, and inversely mapping the partitioning result into a final target grid, and accordingly determining an optimal load balancing strategy, so that the efficiency of determining the load balancing strategy can be improved.

Requests are received to write data into extents. The extents are logically contiguous address ranges having corresponding sizes that are larger than a minimum atomic write size of a storage system. A determination as to whether an amount of data written to a particular extent of the plurality of extents has reached a threshold is made. In response to determining that the amount of data written to the particular extent has reached the threshold, data written to the particular extent is stored in one or more managed flash storage devices utilizing a flash translation layer organized to store blocks of a size corresponding to the size of the extents.

Systems and methods for intelligent processing of requests in an artificial intelligence system, including receiving an incoming request at an AI broker, evaluating the incoming request to determine routing characteristics, selecting selected specialized language models from a plurality of specialized language models based on the routing characteristics, computing performance metrics for the one or more selected specialized language models, routing the incoming request to the selected specialized language models based on the performance metrics, and generating a final result by receiving and processing results from the selected specialized language models.

1. Name of the product of this design: Main body of supercomputer. 2. Purpose of the product of this design: The supercomputer is used for data storage and processing, and the main body is used to accommodate internal components. 3. The key point of the design of this product lies in its shape. 4. The picture or photo that best illustrates the design points: Stereoscopic drawing 1. 5. Other situations that require explanation: The structure indicated by the dotted line is limited to the structure that does not require protection.

The invention belongs to the field of molecular dynamics, and provides a molecular dynamics adjacency list construction optimization method, a molecular dynamics adjacency list construction optimization system and a super computer platform in order to solve the problem that the cache hit rate is reduced when a traditional adjacency list is constructed. The molecular dynamics adjacency list construction optimization method comprises the following steps: mapping atomic coordinates in each physical space to three-dimensional integer grid coordinates; carrying out one-dimensional mapping coding on the three-dimensional integer grid coordinates of all the atoms and sorting all the atoms; clustering all the sequenced atoms to obtain a plurality of atom clusters and initializing sharing states in the atom clusters; constructing a candidate neighbor region of each atomic cluster based on the initialized sharing state in each atomic cluster; and uniformly traversing the candidate neighbor regions of each atomic cluster according to a vectorization instruction, constructing an adjacency list of each atom of each atomic cluster, and writing the adjacency list into a global adjacency list, so that larger-scale and higher-efficiency molecular simulation is realized.

The application discloses a low-energy-consumption passive heat dissipation system for supercomputing / data centers, wherein the top end of each type of containerized extensible case is connected to one end of a non-activated cooler through a pipeline, the other end of the non-activated cooler is connected to one end of a liquid distribution system through a pipeline, and the other end of the liquid distribution system is connected to the bottom end of each type of containerized extensible case, so as to form a closed pipeline system. The heat dissipation system adopts an "insulation cooling liquid immersion cooling" + "separated heat pipe" + "high-efficiency compact condenser" heat and mass transfer process, is of passive heat dissipation throughout, greatly reduces the heat dissipation power consumption of the supercomputing / data center, and can realize an ultra-low PUE value. In addition, the system can realize higher compactness of architecture assembly and reduce the space occupation scale of the supercomputing / data center. The heat dissipation system has high heat dissipation power, high heat dissipation capacity expandability, fast and convenient computer on / off operation, energy saving and environmental protection.

Systems and methods of answering queries using one or more h-LLMs including receiving user prompts via an API at a query layer from a user interface, determining an analysis mode of the user prompts to be a real-time mode, transmitting a query prompt including the user prompts from the query layer to a real-time layer responsive to determining the analysis mode being a real-time mode, the real-time layer including an h-LLM and being capable of generating a result in a response period on the order of seconds, receiving a real-time response from the real-time layer responsive to the query prompt, and transmitting the real-time response to the user interface via the API.

A large-model training optimization method for domestic supercomputer systems is applied to computing devices in multiple domestic supercomputer systems. Each computing device is equipped with a GPU, and the GPU contains at least one process. The method is applied to the Megatron-DeepSpeed ​​framework and includes: determining the processes required for large-model training, and determining the process group to which each process belongs; based on the order of tensor parallelism, pipeline parallelism, and data parallelism in the Megatron-DeepSpeed ​​framework, simultaneously constructing process groups, each of which includes at least one process; each process performs multiple forward and reverse calculations in the parallel training framework, and data exchange and synchronization are performed through the process group's communication mechanism. The forward and reverse calculations include collective communication. This method can improve the training efficiency of large-model training on domestic supercomputers.

The air-liquid homologous indirect evaporative cooling all-in-one unit suitable for the supercomputing center has the advantages that cold air and cold water are produced at the same time, energy is saved, consumption is reduced, safety and reliability are achieved, and low-noise operation is achieved. The indirect evaporative cooling air conditioning unit module and the indirect evaporative cooling water chilling unit module are organically combined and share one outdoor side exhaust fan, so that the number of equipment and the operation energy consumption are reduced, and the initial investment and maintenance cost are reduced; a dry channel in the air conditioning unit module is designed to be positive pressure, a wet channel is designed to be negative pressure, leaked water is effectively prevented from entering a machine room, and the system safety is improved; and a low-frequency axial flow fan is adopted as an exhaust fan, so that noise pollution is remarkably reduced. The unit realizes accurate cooling of different heating elements of the supercomputer center, optimizes the system energy efficiency, meets the green and low-carbon requirements, and solves the problems of uncoordinated equipment, high energy consumption, loud noise and the like in the prior art.

Requests are received to write data into extents. The extents are logically contiguous address ranges having corresponding sizes that are larger than a minimum atomic write size of a storage system. A determination as to whether an amount of data written to a particular extent of the plurality of extents has reached a threshold is made. In response to determining that the amount of data written to the particular extent has reached the threshold, data written to the particular extent is stored in one or more managed flash storage devices utilizing a flash translation layer organized to store blocks of a size corresponding to the size of the extents.

The ever increasing number of orbiting bodies in low Earth orbit has made it infeasible to calculate potential conjunctions between orbiting bodies more than a few days in advance, even with the aid of supercomputers. Disclosed embodiments utilize machine learning to predict potential conjunctions between orbiting bodies faster than state-of-the art systems by orders of magnitude. This enables potential conjunctions to be identified well in advance (e.g., 30 days or more), so that they may be prioritized (e.g., for fine calculations), visualized, and remediated (e.g., via control of the impacted satellites).

The application provides a remote cooperative post-processing method and device for supercomputers, equipment and medium. The method comprises the following steps: determining a target quantity participating in post-processing, a post-processing type and associated configuration information; sending the target quantity, the post-processing type and the associated configuration information to a supercomputer cluster, so that the supercomputer cluster reads corresponding simulation data from a matched simulation data file in parallel according to the target quantity, performs compression processing on the read simulation data to obtain to-be-rendered data, sends the to-be-rendered data to a client if the data quantity of the to-be-rendered data is less than a threshold, performs rendering on the to-be-rendered data according to the post-processing type and the associated configuration information to obtain graphic rendering data and sends the graphic rendering data to the client if the data quantity of the to-be-rendered data is greater than or equal to the threshold. The application aims to solve the problems of insufficient local processing capacity and unsmooth interface display.

A resource management mechanism oriented to distributed multi-quantum equipment comprises quantum computing resource virtualization, an independent QPU equipment topological structure in a distributed quantum computing environment and a connection relation between QPUs are abstracted into a graph data structure, and updating is carried out according to the real-time state of the quantum equipment. When real quantum computing resources in the system are insufficient, a super computer is used for simulating execution of a quantum circuit; a line scheduling mechanism in a distributed quantum computing environment determines whether a quantum line needs to be cut based on a quantum device state in a system and a resource demand of the quantum line to be executed. When distributed quantum computing needs to be executed, a popularity / weak connectivity mapping optimization algorithm is adopted, and the fidelity of a quantum program is improved. According to the invention, a brand new solution is provided for managing quantum computing resources in a distributed computing environment, the utilization rate of the quantum computing resources is improved, and the execution fidelity of a quantum circuit is improved.

Systems and methods are provided for implementing encryption of data-in-motion and / or otherwise stored data using a key server and a secure enclave of a Network Interface Card (NIC). The NIC acts as a passthrough between the client device and the shared infrastructure of the supercomputer system to help ensure data security in a massively scaled and distributed system. For example, in response to an enrollment process that stores a decrypted key in the secure enclave of a NIC, the NIC can receive a data packet from a client device. The NIC can transmit a key request to a key server that includes an encrypted key corresponding to the decrypted key. The key server can look up the previously stored private / public key pair to authenticate the NIC. The key server can provide private / public key pair to the NIC to allow the NIC to later encrypt data-in-motion.