FFT DSP Engine with Feedback Architecture for Low-Latency Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing integrated circuit devices struggle to efficiently perform calculations for both digital signal processing and machine learning applications, as circuitry optimized for one domain is often not well-suited for the other.

Innovation Solution

A digital signal processing (DSP) block is designed to perform multiple Fast Fourier Transform (FFT) and inverse FFT operations, incorporating a feedback structure to reduce latency and power consumption, and including both fixed and variable FFT engines along with a scale/offset block.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If circuitry is optimized for digital signal processing, then DSP performance is improved, but machine learning performance deteriorates

Engineering Contradiction:
ImproveDSP processing speedVSAvoidMachine learning capability
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent implements a unified hardware architecture that can perform both FFT operations for digital signal processing and matrix multiplication operations for machine learning. The same computational units (multipliers, accumulators, data paths) are configured to execute different algorithms based on control signals, eliminating the need for separate specialized circuitry for each domain while maintaining high performance in both applications.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Measurement precision

If FFT operations are performed with data read/write from hardware accelerator, then processing accuracy is maintained, but latency and power consumption increase

Engineering Contradiction:
ImproveProcessing accuracyVSAvoidLatency
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent implements a feedback structure where intermediate results from FFT operations are retained within the hardware accelerator's internal memory or registers. This allows subsequent FFT operations to reuse previously computed data without requiring external read/write operations, thereby reducing latency and power consumption while maintaining processing accuracy through preserved computational precision.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If FFT operations are performed with data read/write from hardware accelerator, then data freshness is ensured, but power consumption increases

Engineering Contradiction:
ImproveData freshnessVSAvoidPower consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent employs a feedback mechanism where the hardware accelerator monitors operational patterns and retains data in internal storage when subsequent operations require the same data. This feedback-controlled data retention strategy reduces frequent external memory access, thereby lowering power consumption while ensuring data freshness is maintained through controlled caching of intermediate results.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS12278630B2Fast fourier transform (FFT) based digital signal processing (DSP) engine
Publication Date: 2025.04.15 ALTERA CORP
  • US12278630B2 patent drawing
  • US12278630B2 patent drawing
  • US12278630B2 patent drawing

AI summary

A digital signal processing (DSP) block includes a Fast Fourier Transform (FFT) unit capable of performing an FFT operation. The FFT unit includes a first FFT engine capable of converting a signal between a time-domain and a frequency-domain and the first FFT engine is a fixed size FFT engine. The FFT unit also includes a second FFT engine communicatively coupled to the first FFT engine and the second FFT engine is a variable size FFT engine. The FFT unit also includes a scale/offset block communicatively coupled to the second FFT engine and the scale/offset block is capable of performing a multiplication operation, an addition operation, or a combination thereof on an output of the second FFT engine.