A method for reducing energy consumption of a large-scale distributed machine learning system
A machine learning and distributed technology, applied in instruments, resource allocation, energy-saving computing, etc., can solve problems such as server power waste, overall performance degradation, and long iteration time, so as to reduce system energy consumption, improve utilization rate, and shorten execution time. the effect of time
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Example Embodiment
[0020] The method for reducing energy consumption of large-scale distributed machine learning proposed by the present invention has the following steps:
[0021] Step 1: The scheduler collects the real-time information of the CPU, GPU, memory, and disk I / O of the working machine and sends it to the state memory.
[0022] Step 2: The state memory uses the received real-time information of the processor, memory, and disk I / O to calculate the workload status of the workload (CPU usage, GPU usage, memory usage, and disk I / O usage).
[0023] Step 3: The scheduling policy manager reads the load information on the state memory. The load status of different workloads at the same time is used to predict the load type of machine learning tasks (computation-intensive, I / O-intensive, GPU-accelerated, hybrid), and the load curve at different moments is used to predict the workload of a period of time in the future load.
[0024] Step 4: When the machine learning task arrives, first use the schedu...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap