Wireless Baseband Modem Datapath With Parallel Compute-In-Memory Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional modem signal processing functions are energy inefficient due to separate computation and memory, leading to high energy consumption and reduced throughput.
Innovation Solution
Implementing a compute-in-memory (CIM) system that integrates processing elements within memory chips, allowing parallel data access and processing, thereby eliminating row-by-row memory access latency and reducing active time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If separate computation and memory are used with row-by-row memory access, then data processing can be performed, but energy consumption increases and throughput decreases
Solution Approach 1:
The patent merges computation and memory into a unified compute-in-memory architecture where processing elements are integrated directly within the memory chip. This eliminates the need for separate memory access operations, allowing data to be processed in-place without being transferred between memory and computation units, thereby simultaneously improving throughput and reducing energy consumption.
Solution Approach 2:
The patent introduces processing elements as intermediary components embedded within the memory structure. These processing elements act as mediators that can perform computations directly on the stored data without requiring data to be moved to separate computation units, thus eliminating the energy-wasting row-by-row access pattern while maintaining high throughput.
2Speed
If separate computation and memory are used with row-by-row memory access, then data processing can be performed, but processing latency increases
Solution Approach 1:
By merging computation and memory into a single integrated structure, the patent eliminates the time-consuming data transfer between separate memory and computation units. Processing elements embedded within the memory chip can immediately access and process data in-place, dramatically reducing memory access latency and improving overall processing speed.
Solution Approach 2:
The embedded processing elements serve as intermediaries that bridge the gap between storage and computation. They enable direct in-place processing of data within the memory array, eliminating the sequential row-by-row access latency and enabling parallel processing operations that significantly improve processing speed.
3Ease of operation
If conventional separate memory and computation architecture is used, then data can be stored and processed, but the system complexity increases due to separate hardware blocks
Solution Approach 1:
The patent merges previously separate memory and computation hardware blocks into a single integrated compute-in-memory unit. This consolidation reduces the number of discrete components and interconnections required, simplifying the overall hardware architecture while maintaining full data storage and processing functionality.
Solution Approach 2:
The integrated compute-in-memory architecture provides multi-functionality by combining storage and computation capabilities in a single hardware block. This universal structure can perform both data storage and various processing operations without requiring separate dedicated hardware for each function, thereby reducing system complexity and improving ease of operation.
Data Source
AI summary
System and method embodiments are disclosed for merging computing hardware datapath to reduce energy consumption and improve throughput of modem signal processing functions on a Modem. Compute-in-memory (CIM) is used for data access and processing in parallel using processing elements integrated within memory chip. The processing throughput may significantly improve since data is accessed directly from the computational memory storage in parallel for processing and the clock speed for data processing may run at a higher spec than conventional approaches. Energy consumption for Modem processing is also reduced.


