Abstract: Disclosed herein is a method for distributed training of an AI model in a channel-sharing network environment. The method includes determining whether data parallel processing is applied, calculating a computation time and a communication time when input data is evenly distributed across multiple computation devices, and unevenly distributing the input data across the multiple computation devices based on the computation time and the communication time.

Abstract: Disclosed herein are a method for machine-learning parallelization using host CPUs of a multi-socket structure and an apparatus therefor. The method, performed by the apparatus for machine-learning parallelization using host CPUs of a multi-socket structure, includes a compile phase in which a learning model is split at a layer level for respective pipeline stages and allocated to Non-Uniform Memory Access (NUMA) nodes for respective CPU sockets and a runtime phase in which parameters required for learning are initialized and multiple threads generated in consideration of a policy of each parallelism algorithm are executed by being allocated to respective cores included in the NUMA node.

Abstract: A display device includes a bank including an opening exposing a surface of a base. The bank further includes side surfaces adjacent to an upper surface. The side surfaces slope downward from the upper surface toward an opening in an organic film pattern. A plurality of fine holes is formed on the upper surface and the side surfaces, the bank may also include a plurality of inner holes.

Abstract: A pixel includes first, second, and third sub-pixels each including an emission area and a non-emission area. Each of the first, second, and third sub-includes a pixel circuit layer; a first electrode on the pixel circuit layer; a pixel defining layer on the first electrode and including an opening to expose an area of the first electrode; an emission layer on the pixel defining layer; a second electrode on the emission layer; a thin film encapsulation layer over the second electrode; a color filter on the thin film encapsulation layer; and an overcoat layer over the color filter. The overcoat layer has a refractive index greater than a refractive index of the color filter. A color filter of the second sub-pixel overlaps a color filter of each of the first and third sub-pixels in the non-emission area.

Abstract: An embodiment composition for solid electrolyte membranes of all-solid-state batteries includes a sulfide-based solid electrolyte and a cross-linking agent including two or more acrylate functionalities. An embodiment method of manufacturing a solid electrolyte membrane for an all-solid-state battery includes forming a composition including a sulfide-based solid electrolyte and a cross-linking agent including two or more acrylate functionalities and cross-linking the composition.

Abstract: An anodeless all-solid-state battery includes an anode current collector, a composite structure layer positioned on the anode current collector, a solid electrolyte positioned on the composite structure layer, and a cathode positioned on the solid electrolyte, in which the composite structure layer includes a carbon layer including a carbon material, and a metal deposition layer positioned on the carbon layer and including lithiophilic metal particles.

Abstract: Disclosed herein are a hybrid memory management apparatus and method for an many-to-one virtualization environment. The hybrid memory management apparatus is implemented in an inverse-virtualization-based multi-node computing system including multiple physical nodes, each containing hybrid memory in which DRAM and NVRAM coexist, a virtual machine, and hypervisors, and includes memory for storing at least one program, and a processor for executing the program, wherein the program includes a remote request service module for processing a page-related request with reference to the hybrid memory and responding to the page-related request by transmitting a result of processing, an internal request service module for processing an internal page fault request with reference to a hybrid memory and responding to the internal page fault request, and a data arrangement module for responding to an inquiry request for a location at which a newly added page is to be arranged in the hybrid memory.

Abstract: Disclosed herein are an apparatus and method for page allocation in a many-to-one virtualization environment. The method may include determining whether a page fault interrupt is caused by page initialization for page allocation, sending an ownership change message to a node having ownership of the corresponding page when the page fault interrupt is determined to be caused by page initialization, and initializing the corresponding page upon receiving an ownership-change-processing-complete message.

Abstract: Disclosed herein are an apparatus for preloading data in a distributed computing environment and a method using the same. The method includes selecting a local preloading target that each of multiple computers connected over a network is to preload into the local memory thereof, registering a local preloading task corresponding to the local preloading target in local preloading metadata, and asynchronously starting the local preloading task at a preset time based on the local preloading metadata. The local preloading metadata is stored in a page other than the page in which remote preloading metadata for managing a remote preloading task is stored.

Abstract: Disclosed herein are an apparatus and method for processing fast persistent writing to nonvolatile memory. A method for processing fast persistent writing to a nonvolatile memory includes writing log data from a last storage position in a data block created in a certain size, reflecting the entire data block in a storage medium, increasing the last storage position by a size of the written log data, and reflecting the entire data block in a storage medium, wherein the data block is linearly linked to an additional data block in a creation order, and wherein the data block includes a log data storage region, information about a last storage position where the log data is stored in the log data storage region, and information about a linked next data block.

Abstract: Disclosed herein are a synchronization method for improving the concurrent read performance of a critical section in distributed shared memory and an apparatus for the same. The synchronization method, performed by a distributed-shared-memory management apparatus in a physical node of a multi-node system, includes checking whether a lock is held on each node based on a read-write lock having lock variables for respective nodes in a distributed shared memory environment, acquiring a lock for a read operation or a write operation in consideration of whether a lock is held on each node, and releasing the lock based on the lock variables for the respective nodes when the read operation or the write operation is terminated.

Abstract: Disclosed herein are a distributed system and a method for operating the distributed system. The method for operating a distributed system including a server and multiple clients includes acquiring, by a first client of the multiple clients, a lock on a shared resource using a first table of the server and a second table of the client, and releasing, by the first client, a lock on the shared resource using the first table and the second table, wherein the first table is a lock (DSLock) table for storing information about a distributed shared resource, and the second table is a data structure (DSLock_node) table for a lock request.

Abstract: Disclosed herein are a hybrid memory management apparatus and method for an many-to-one virtualization environment. The hybrid memory management apparatus is implemented in an inverse-virtualization-based multi-node computing system including multiple physical nodes, each containing hybrid memory in which DRAM and NVRAM coexist, a virtual machine, and hypervisors, and includes memory for storing at least one program, and a processor for executing the program, wherein the program includes a remote request service module for processing a page-related request with reference to the hybrid memory and responding to the page-related request by transmitting a result of processing, an internal request service module for processing an internal page fault request with reference to a hybrid memory and responding to the internal page fault request, and a data arrangement module for responding to an inquiry request for a location at which a newly added page is to be arranged in the hybrid memory.

Abstract: Disclosed herein are a metadata server and a method for distributing metadata in units of directories using the metadata server. The method includes receiving a request for at least one of creation, deletion and retrieval from a user file system; acquiring an inode corresponding to an identifier of a parent inode, which is included in the request; searching the parent inode for a dentry corresponding to the request when the request is a request to create a directory; selecting a secondary metadata server in which a new directory is to be created in order to distribute metadata in units of directories; requesting the selected secondary metadata server to create an inode; receiving an identifier of the created inode from the secondary metadata server; and adding a new dentry to the parent inode using the identifier of the inode.

Abstract: A distributed storage server includes a plurality of data server devices and a plurality of metadata server devices. The metadata server devices store metadata associated with data which is distributively stored in the data server devices. A selected metadata server device checks whether a dangling directory occurs by performing a rename operation, based on information associated with a full path of a parent directory of a source and a full path of a parent directory of a target included in a request of the rename operation. When it is determined that the dangling directory does not occur, the selected metadata server device processes transactions directed to the metadata such that the rename operation is performed. The checking whether the dangling directory occurs is performed before a transaction period for processing the transactions.

Abstract: Disclosed herein are a distributed system and a method for operating the distributed system. The method for operating a distributed system including a server and multiple clients includes acquiring, by a first client of the multiple clients, a lock on a shared resource using a first table of the server and a second table of the client, and releasing, by the first client, a lock on the shared resource using the first table and the second table, wherein the first table is a lock (DSLock) table for storing information about a distributed shared resource, and the second table is a data structure (DSLock node) table for a lock request.

Abstract: Disclosed herein is a distributed file system using a torus network. The distributed file system includes multiple servers. The location of a master server may be determined to shorten the latency of data input/output. The location of the master server may be determined such that the distance between the master server and a node farthest away from the master server, among nodes, is minimized. When the location of the master server is determined, the characteristics of the torus network and the features of a propagation transmission scheme may be taken into consideration.

Abstract: Disclosed herein are a method for automatically allocating IP addresses for a distributed storage system in a large-scale torus network and an apparatus for the method. The method for automatically allocating IP addresses includes acquiring neighboring-node information from multiple storage nodes that constitute the system; generating torus network topology for the multiple storage nodes by combining the neighboring-node information; generating IP address information through which a position of each of the multiple storage nodes in a structure corresponding to the torus network topology can be identified; and allocating IP addresses to the multiple storage nodes by delivering the IP address information to the multiple storage nodes.

Abstract: Disclosed herein are a method and system for supporting user-level DMA I/O in a distributed file system environment. The system includes a DMA I/O library located in a user space and configured to process data-processing requests of an application using a control channel or a data channel, a distributed file system client process located in the user space and configured to process data while communicating with a remote distributed storage cluster, a VFS located in a kernel space, and a FUSE kernel module located in the kernel space, wherein the control channel is a path along which the DMA I/O library and the distributed file system client process are connected to each other through the VFS and the FUSE kernel module, and wherein the data channel is a path along which the DMA I/O library and the distributed file system client process are directly connected to each other.

Abstract: A distributed file system includes metadata servers and data servers. The metadata server includes a selecting unit to select a data server from the data servers, a chunk allocation requesting unit to request that the selected data server perform chunk allocation, a chunk list managing unit to insert a list of chunks transmitted from the data server into a chunk list pool and determine an arrangement method of the chunk list pool, and a chunk fetching unit to fetch available chunk information from the chunk list pool. The data server includes a receiving unit to receive request for chunk allocation from the metadata server, a chunk allocating unit to allocate chunks in response to the request for chunk allocation and write a list of chunks based on information about the allocated chunk, and a transmitting unit to transmit the list of chunks to the metadata server.