17994817. IN-MEMORY COMPUTATION SYSTEM WITH COMPACT STORAGE OF SIGNED COMPUTATIONAL WEIGHT DATA simplified abstract (STMICROELECTRONICS S.r.l.)
Contents
- 1 IN-MEMORY COMPUTATION SYSTEM WITH COMPACT STORAGE OF SIGNED COMPUTATIONAL WEIGHT DATA
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 IN-MEMORY COMPUTATION SYSTEM WITH COMPACT STORAGE OF SIGNED COMPUTATIONAL WEIGHT DATA - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Potential Commercial Applications
- 1.9 Possible Prior Art
- 1.10 Original Abstract Submitted
IN-MEMORY COMPUTATION SYSTEM WITH COMPACT STORAGE OF SIGNED COMPUTATIONAL WEIGHT DATA
Organization Name
Inventor(s)
Marcella Carissimi of Treviolo (IT)
Paolo Sergio Zambotti of Milano (IT)
Riccardo Zurla of Binasco (MI) (IT)
IN-MEMORY COMPUTATION SYSTEM WITH COMPACT STORAGE OF SIGNED COMPUTATIONAL WEIGHT DATA - A simplified explanation of the abstract
This abstract first appeared for US patent application 17994817 titled 'IN-MEMORY COMPUTATION SYSTEM WITH COMPACT STORAGE OF SIGNED COMPUTATIONAL WEIGHT DATA
Simplified Explanation
The IMC circuit described in the patent application consists of memory cells arranged in a matrix, with computational weights stored in groups of cells. Each row of groups of cells includes a positive and negative word line, while each column includes a bit line. The IMC operation involves applying a word line signal to the positive/negative word line of the group of cells based on the sign of the coefficient data, resulting in a positive MAC output on the bit line. In a second step, a word line signal is applied to the negative/positive word line of the group of cells, leading to a negative MAC output on the bit line. The final IMC operation result is obtained by calculating the difference between the positive and negative MAC operations.
- Memory cells arranged in a matrix
- Computational weights stored in groups of cells
- Positive and negative word lines in each row
- Bit lines in each column
- Positive and negative MAC operations based on coefficient data
Potential Applications
The technology described in the patent application could be applied in various fields such as:
- Artificial intelligence
- Machine learning
- Signal processing
Problems Solved
This technology helps in:
- Efficient computation of MAC operations
- Optimizing memory cell organization
- Enhancing computational performance
Benefits
The benefits of this technology include:
- Faster processing speeds
- Improved accuracy in calculations
- Reduced power consumption
Potential Commercial Applications
Potential commercial applications of this technology could include:
- Neural network processors
- Image recognition systems
- Data analysis tools
Possible Prior Art
One possible prior art related to this technology is the use of memory cells in matrix organization for computational operations. This concept has been explored in various research studies and patents in the field of artificial intelligence and signal processing.
Unanswered Questions
How does this technology compare to existing MAC operation methods?
This article does not provide a direct comparison between this technology and traditional methods of performing MAC operations. Further research or testing may be needed to evaluate the efficiency and effectiveness of this new approach.
What are the potential limitations of implementing this technology in practical applications?
The article does not address any potential challenges or limitations that may arise when implementing this technology in real-world scenarios. It would be important to consider factors such as scalability, cost, and compatibility with existing systems.
Original Abstract Submitted
An IMC circuit includes a memory cells arranged in matrix. Computational weights for an IMC operation are stored in groups of cells. Each row of groups of cells includes a positive and negative word linen. Each column of groups of cells includes a bit line. The IMC operation includes a first elaboration where a word line signal is applied to the positive/negative word line of the group of cells depending on the positive/negative sign, respectively, of the coefficient data, with a positive MAC output on the bit line. In a second elaboration, a word line signal is applied to the negative/positive word line of the group of cells depending on the positive/negative sign, respectively, of the coefficient data, with a negative MAC output on the bit line. The IMC operation result is obtained from a difference between the positive and negative MAC operations.
