Accelerator Kernel Assignment for Database Scan Priority
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing database query systems face inefficiencies in allocating hardware resources for table scans, particularly when multiple queries are processed in parallel, leading to suboptimal use of accelerator resources.
Innovation Solution
An accelerator determines priority values for each scan based on cache statistics and table size, assigning kernels to scans proportionally, and manages kernel allocation to ensure efficient use of hardware resources, including cache flushing and fallback to software scans when appropriate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If accelerator resources are allocated to handle multiple scans in parallel, then processing throughput is improved, but resource allocation efficiency deteriorates due to lack of priority-based distribution
Solution Approach 1:
The system changes the parameter of resource allocation from uniform distribution to priority-based distribution. By calculating scan priorities based on cache statistics and dynamically adjusting kernel allocation proportions, the system optimizes resource utilization while maintaining high processing throughput for multiple parallel scans.
2Ease of operation
If kernels are assigned to scans without priority consideration, then resource utilization is simplified, but processing efficiency deteriorates due to unequal scan importance
Solution Approach 1:
The system implements dynamic kernel assignment where the number of kernels assigned to each scan is not fixed but adjusts based on calculated priorities. The kernel assignment unit dynamically determines allocation proportions according to scan priorities, allowing the system to adapt to varying scan importance while maintaining operational simplicity through automated calculations.
3Quantity of substance
If software scans are used instead of hardware accelerators, then resource requirements are reduced, but processing speed deteriorates
Solution Approach 1:
The system applies different processing qualities to different scans based on their priorities and data characteristics. High-priority scans with large data volumes that benefit from hardware acceleration receive kernel assignments, while lower-priority scans may use software implementations. This localized quality differentiation optimizes overall system performance by matching resource intensity to actual need.
Data Source
AI summary
An apparatus is disclosed. The apparatus may include a storage device, which may store a database including a table. The apparatus may also include an accelerator connected to the storage device. The accelerator may include a kernel. A scan, associated with a query, may access data from the table in the database stored on the storage device. A scan priority calculator may calculate a priority of the scan. A kernel assignment unit may assign the kernel to the scan based at least in part on the priority of the scan.