Abstract: An electronic apparatus for performing machine learning a method of machine learning, and a non-transitory computer-readable recording medium are provided. The electronic apparatus includes an operation module configured to include a plurality of processing elements arranged in a predetermined pattern and share data between the plurality of processing elements which are adjacent to each other to perform an operation; and a processor configured to control the operation module to perform a convolution operation by applying a filter to input data, wherein the processor controls the operation module to perform the convolution operation by inputting each of a plurality of elements configuring a two-dimensional filter to the plurality of processing elements in a predetermined order and sequentially applying the plurality of elements to the input data.

Abstract: An operating method of a computing processor includes: performing a computing operation on a data unit assigned from among a plurality of data units that are partitioned from multi-media data; obtaining operation completion information of the computing operation performed on the data unit from among the plurality of data units, from at least one predetermined computing processor from among a plurality of computing processors; storing the operation completion information obtained from the at least one predetermined computing processor; and determining whether computing operations performed on the multi-media data have been completed, based on the operation completion information obtained from the at least one predetermined computing processor and stored, and the computing operation performed on the assigned data unit.

Abstract: A computing system is disclosed. The computing system according to one embodiment of the present disclosure comprises: a memory device for storing an application program; a processor for executing a loader for loading data of the application program into a memory space allocated for execution of the application program; a local memory having a width corresponding to the size of a register of the processor; and a constant memory having a width smaller than that of the local memory, wherein, according to the size of constant data included in the application program, the processor loads the constant data into one of the local memory and the constant memory.

Abstract: An operating method of a computing processor includes: performing a computing operation on a data unit assigned from among a plurality of data units that are partitioned from multi-media data; obtaining operation completion information of the computing operation performed on the data unit from among the plurality of data units, from at least one predetermined computing processor from among a plurality of computing processors; storing the operation completion information obtained from the at least one predetermined computing processor; and determining whether computing operations performed on the multi-media data have been completed, based on the operation completion information obtained from the at least one predetermined computing processor and stored, and the computing operation performed on the assigned data unit.

Abstract: A processor and a control method thereof are processed. The processor includes an instruction fetch module configured to receive a first instruction of an interrupt service routine without backup of data stored in a register in response to processing of the interrupt service routine being requested, a detecting module configured to analyze the received first instruction to determine whether the data stored in the register needs to be changed, an instruction generating module configured to generate a second instruction for storing data in a temporary memory when the stored data is initially changed, an instruction selecting module configured to sequentially select the generated second instruction and first instruction; and a control module configured to perform the second instruction and the first instruction.

Abstract: A computing system is disclosed. The computing system according to one embodiment of the present disclosure comprises: a memory device for storing an application program; a processor for executing a loader for loading data of the application program into a memory space allocated for execution of the application program; a local memory having a width corresponding to the size of a register of the processor; and a constant memory having a width smaller than that of the local memory, wherein, according to the size of constant data included in the application program, the processor loads the constant data into one of the local memory and the constant memory.

Abstract: An electronic apparatus for performing machine learning a method of machine learning, and a non-transitory computer-readable recording medium are provided. The electronic apparatus includes an operation module configured to include a plurality of processing elements arranged in a predetermined pattern and share data between the plurality of processing elements which are adjacent to each other to perform an operation; and a processor configured to control the operation module to perform a convolution operation by applying a filter to input data, wherein the processor controls the operation module to perform the convolution operation by inputting each of a plurality of elements configuring a two-dimensional filter to the plurality of processing elements in a predetermined order and sequentially applying the plurality of elements to the input data.

Abstract: A processor and a control method thereof are processed. The processor includes an instruction fetch module configured to receive a first instruction of an interrupt service routine without backup of data stored in a register in response to processing of the interrupt service routine being requested, a detecting module configured to analyze the received first instruction to determine whether the data stored in the register needs to be changed, an instruction generating module configured to generate a second instruction for storing data in a temporary memory when the stored data is initially changed, an instruction selecting module configured to sequentially select the generated second instruction and first instruction; and a control module configured to perform the second instruction and the first instruction.

Abstract: A method of generating a code for a remote procedure call (RPC) includes obtaining a source code including information indicating a part where the RPC is to be performed, and generating a code for calling the RPC and a code for executing an RPC procedure, by analyzing the source code including information indicating the part where the RPC is to be performed.

Abstract: A modulo scheduling method including calculating at least two candidate initiation intervals between adjacent iterations, searching for schedules of the instructions in parallel by using the candidate initiation intervals, and selecting a schedule determined to be valid from among the searched schedules.

Abstract: A method of generating a code for a remote procedure call (RPC) includes obtaining a source code including information indicating a part where the RPC is to be performed, and generating a code for calling the RPC and a code for executing an RPC procedure, by analyzing the source code including information indicating the part where the RPC is to be performed.

Abstract: A method of measuring software performance includes inserting a performance measurement code into a source code, stalling a target system, on which the code is executed by a processor, and a performance counter based on the performance measurement code, transmitting performance data corresponding to a stalled time point when the target system and the performance counter are stalled to a host system configured to store the performance data corresponding to the stalled time point, and resuming execution of the source code by the target system and of the performance counter while the performance data is transmitted and stored.

Abstract: Provided are a storage device and a method of distributed processing of multimedia data. In the method, the storage device stores multimedia data, initiates an interface configured to share data between a host and the storage device, receives a multimedia data request from the host, processes the multimedia data based on the received multimedia data request, and transmits the processed multimedia data to the host through the interface.

Abstract: A modulo scheduling method including calculating at least two candidate initiation intervals between adjacent iterations, searching for schedules of the instructions in parallel by using the candidate initiation intervals, and selecting a schedule determined to be valid from among the searched schedules.

Abstract: Methods and apparatuses are provided for converting data efficiently and for converting input data by automatically calculating a size of an input buffer in accordance with information regarding input data when the input data is converted. The method of converting data may include receiving size information of an output buffer, receiving range information of input data used to calculate output data corresponding to each point within the output buffer, calculating size information of an input buffer using the size information of the output buffer and the range information of the input data, loading the input data using the size information of the input buffer, and calculating the output data using the loaded input data.

Abstract: Provided are a load balancing method and a load balancing apparatus in a symmetric multi-processor system. The load balancing method includes selecting at least two processors based on a load between a plurality of processors, from among the plurality of processors, migrating a predetermined task stored in a run queue of a first processor to a migration queue of a second processor, and migrating the predetermined task stored in the migration queue of the second processor to a run queue of the second processor. Accordingly, a run queue of a processor is not blocked while migrating a task, an immediate response of the run queue is possible, and a waiting time of a scheduler is reduced. Consequently, the scheduler can speedily perform context switching, and thus performance of the entire operating system is improved.

Abstract: Provided are an apparatus for reconfiguring a mapping method and a scheduling method in a reconfigurable multi-processor system. A single function is mapped to a reconfigurable processor. When a task is created in the reconfigurable multi-processor system, a function of the task is dynamically mapped to a host processor or a reconfigurable processor, thereby removing temporal sharing between functions on the reconfigurable processor and thus reducing the number of times reconfiguration is performed. The overhead of the reconfigurable processor is minimized and the reconfigurable processor is optimized for a dynamic multi-application environment.

Abstract: A method, medium and apparatus for storing and restoring a register context for a fast context switching between tasks is disclosed. The method, medium and apparatus may improve overall operating speed of a system by increasing the speed of context switching. The method may include adding an update code for updating information of live registers to a task file that includes a code of a task to perform a specified function, converting the task file having the update code added thereto into a run file, updating the information of the live registers with the update code during running of the task using the run file, and storing a live register context according to the updated information of the registers.

Abstract: A multi-tasking method performs a plurality of tasks according to priority of each of the plurality of tasks. It is determined whether a resource that is to be used by a current task is being used by another task, a priority of the current task is compared to a priority of the task that is using the resource according to a result of the determination, and according to a result of the comparison, the priority of the task that is using the resource is increased to a highest priority of priorities of all tasks trying to occupy a CPU. Accordingly, conventional problems in the BPI and IIP schemes can be solved.

Abstract: Provided is a method of managing a configuration memory of reconfigurable hardware which can reconfigure hardware according to hardware configuration information. The method includes: determining at least one slot capable of currently storing the hardware configuration information on the basis of the states of a plurality of slots of the configuration memory; and storing hardware configuration information, which is stored in an external memory, in the determined at least one slot capable of currently storing the hardware configuration information. Accordingly, memory utilization can be improved even in dynamic environment such as data dependent control flow or multi-tasking.