17850090. PERFORMING DISTRIBUTED PROCESSING USING DISTRIBUTED MEMORY simplified abstract (Intel Corporation)
PERFORMING DISTRIBUTED PROCESSING USING DISTRIBUTED MEMORY
Organization Name
Inventor(s)
Abhishek Anil Sharma of Portland OR (US)
Pushkar Ranade of San Jose CA (US)
Sagar Suthram of Portland OR (US)
Wilfred Gomes of Portland OR (US)
Rajabali Koduri of Saratoga CA (US)
PERFORMING DISTRIBUTED PROCESSING USING DISTRIBUTED MEMORY - A simplified explanation of the abstract
This abstract first appeared for US patent application 17850090 titled 'PERFORMING DISTRIBUTED PROCESSING USING DISTRIBUTED MEMORY
Simplified Explanation
The abstract of this patent application describes an apparatus that includes multiple banks to store data and multiple interconnects to connect these banks. The apparatus allows for direct data transfer between two banks in response to a data movement command.
- The apparatus consists of multiple banks and interconnects.
- Each interconnect connects a pair of banks.
- When a data movement command is received, the first bank sends data directly to the second bank through the corresponding interconnect.
Potential Applications
This technology can be applied in various fields where efficient data transfer between banks is required, such as:
- High-performance computing systems
- Data centers
- Networking equipment
- Graphics processing units (GPUs)
- Artificial intelligence and machine learning systems
Problems Solved
This technology addresses the following problems:
- Slow data transfer between banks in traditional systems.
- Bottlenecks in data movement within a system.
- Inefficient utilization of interconnects in data transfer.
Benefits
The benefits of this technology include:
- Faster data transfer between banks.
- Improved system performance and efficiency.
- Reduced latency in data movement.
- Enhanced utilization of interconnects for optimized data transfer.
Original Abstract Submitted
In one embodiment, an apparatus comprises: a plurality of banks to store data; and a plurality of interconnects, each of the plurality of interconnects to couple a pair of the plurality of banks. In response to a data movement command, a first bank of the plurality of banks is to send data directly to a second bank of the plurality of banks via a first interconnect of the plurality of interconnects. Other embodiments are described and claimed.
