An efficient and shuffled unified dye array multi-task parallel processing method

Abstract

The present invention relates to an efficient and disorderly unified dyeing array multi-task parallel processing method, and the present invention comprises the following steps: 1) task input: input the tasks processed by the unified dyeing array; 2) task scheduling; performing task analysis and extracting task features; 3) task execution; 3.1) counting task execution status; 3.2) sending tasks to dyeing array execution units; 4) task waiting; 4.1) counting task waiting status; 4.2) task waiting queue; 5) outputting the tasks processed by the unified dyeing array. The present invention realizes the disorderly parallel processing of multiple tasks, and achieves the goal of improving the processing performance of multi-task complex scenes.

Description

Technical Field

[0001] The invention belongs to the field of computer graphics, and in particular relates to an efficient and shuffleable unified dyeing array multi-task parallel processing method. Background Art

[0002] Currently, unified color arrays are widely used in GPUs to complete task processing. The current application scenarios are becoming more and more complex. In a scenario, not only is the amount of tasks huge, but also it contains multiple types of tasks, which has higher requirements for the unified color array's ability to efficiently process tasks. Currently, there is no efficient and reorderable unified color array multi-task parallel processing method in public information. Summary of the invention

[0003] In order to solve the above-mentioned technical problems existing in the background technology, the present invention provides an efficient and shuffled unified dyeing array multi-task parallel processing method, which realizes the shuffled parallel processing of multiple tasks and achieves the goal of improving the processing performance of multi-task complex scenes.

[0004] The technical solution of the present invention is: the present invention is an efficient and shuffled unified dye array multi-task parallel processing method, the special feature of which is that the method comprises the following steps:

[0005] 1) Task input: input the task to be processed by the unified dye array;

[0006] 2) Task scheduling: perform task analysis, extract task features, perform correlation detection between the currently scheduled task and the tasks in step 3) task execution and step 4) task waiting, determine whether the tasks in step 3) task execution and step 4) task waiting conflict with the currently scheduled task, and perform multi-task parallel scheduling based on the conflict results;

[0007] 3) Task execution;

[0008] 3.1) Statistics of task execution status;

[0009] 3.2) The task is sent to the dye array execution unit;

[0010] 4) Task waiting;

[0011] 4.1) Statistics of task waiting status;

[0012] 4.2) Task waiting queue;

[0013] 5) Output the tasks processed by the unified dye array.

[0014] Preferably, there are two results of task scheduling in step 2). If there is no task conflict, the task is sent to step 3) for task execution. If there is a task conflict, the task is sent to step 4) for task waiting.

[0015] Preferably, in step 2), if the currently scheduled task conflicts with the task being executed in step 3) or conflicts with both the task being executed in step 3) and the task being waited for in step 4), then it is added to the end of the task waiting queue in step 4); if the currently scheduled task conflicts only with the task being waited for in step 4), then it is added to the head of the task waiting queue in step 4).

[0016] Preferably, the task scheduling in step 2) gives priority to the tasks in the task waiting queue in step 4).

[0017] Preferably, the task execution status in step 3.1) includes task characteristics and the situation of task occupying dye array resources.

[0018] Preferably, in step 3.2), tasks that are unrelated to each other are sent to the dye array execution unit.

[0019] Preferably, in step 4.1), the task waiting state includes task characteristics and task occupancy dye array resources.

[0020] Preferably, in step 4.2), the tasks that have not been successfully issued are temporarily stored in a task waiting queue.

[0021] The present invention provides an efficient and shuffled unified dye array multi-task parallel processing method, which inputs tasks, performs task analysis on task scheduling, extracts task features, performs task correlation detection, and realizes shuffled parallel processing of multiple tasks, thereby achieving the goal of improving the processing performance of multiple tasks in complex scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a method module diagram of the present invention; DETAILED DESCRIPTION

[0023] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0024] See also Figure 1 , an efficient and shuffled unified dyeing array multi-task parallel processing method according to an embodiment of the present invention comprises the following steps:

[0025] 1) Task input: input the task to be processed by the unified dye array;

[0026] 2) Task scheduling: perform task analysis, extract task features, perform correlation detection between the currently scheduled task and the tasks in step 3) task execution and step 4) task waiting, determine whether the tasks in step 3) task execution and step 4) task waiting conflict with the currently scheduled task, and perform multi-task parallel scheduling based on the conflict results;

[0027] There are two results of task scheduling. If there is no task conflict, the task is sent to step 3) task execution. If there is a task conflict, it goes to step 4) task waiting.

[0028] If the currently scheduled task conflicts with the task being executed in step 3) or conflicts with both the task being executed in step 3) and the task being waited for in step 4), then it is added to the end of the queue for waiting for the task in step 4); if the currently scheduled task conflicts only with the task being waited for in step 4), then it is added to the head of the queue for waiting for the task in step 4);

[0029] Task scheduling prioritizes the tasks in step 4) the task waiting queue.

[0030] 3) Task execution;

[0031] 3.1) Statistics of task execution status;

[0032] Before tasks are sent to the dye array execution unit, in order to enable the dye array to efficiently process multiple tasks in parallel in a disordered manner, it is necessary to perform task execution status statistics, including task characteristics and the task occupation of dye array resources;

[0033] 3.2) Tasks are sent to the dye array execution unit; tasks that are unrelated to each other are sent to the dye array execution unit;

[0034] 4) Task waiting;

[0035] 4.1) Statistics of task waiting status;

[0036] 4.2) Task waiting queue;

[0037] When a task is not successfully issued, the waiting task will be temporarily stored in the task waiting queue. At the same time, in order for the dyeing array to continue to process multiple tasks in parallel in an efficient and disorderly manner, it is necessary to perform task waiting status statistics, including task characteristics and task occupation of dyeing array resources;

[0038] Temporarily store the tasks that have not been successfully issued in the task waiting queue;

[0039] 5) Output the tasks processed by the unified dye array.

[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An efficient and shuffled unified dye array multi-task parallel processing method, characterized by: The method comprises the following steps: 1) Task input: input the task to be processed by the unified dye array; 2) Task scheduling: perform task analysis, extract task features, perform correlation detection between the currently scheduled task and the tasks in step 3) task execution and step 4) task waiting, determine whether the tasks in step 3) task execution and step 4) task waiting conflict with the currently scheduled task, and perform multi-task parallel scheduling based on the conflict results; There are two results of task scheduling. If there is no task conflict, the task is sent to step 3) task execution. If there is a task conflict, it goes to step 4) task waiting. If the currently scheduled task conflicts with the task being executed in step 3) or conflicts with both the task being executed in step 3) and the task being waited for in step 4), then it is added to the end of the queue for waiting for the task in step 4); if the currently scheduled task conflicts only with the task being waited for in step 4), then it is added to the head of the queue for waiting for the task in step 4); The task scheduling in step 2) gives priority to the tasks in the task waiting queue in step 4); 3) Task execution; 3.1) Statistics of task execution status; 3.2) Sending tasks to the dye array execution unit; Sending unrelated tasks to the dye array execution unit; 4) Task waiting; 4.1) Counting task waiting status; task waiting status includes task characteristics and task occupation of dye array resources; 4.2) Task waiting queue: temporarily store the tasks that have not been successfully issued in the task waiting queue; 5) Output the tasks processed by the unified dye array.

2. The efficient and reorderable unified dyeing array multi-task parallel processing method according to claim 1 is characterized by: The task execution status in step 3.1) includes task characteristics and the situation of task occupying dye array resources.

Images