Configurable DMA Device for Data Transformation and Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional Direct Memory Access (DMA) devices lack the ability to perform complex transformations and searches during memory transfer, requiring additional processing steps by the programmer, which increases computational expense and complexity.
Innovation Solution
A DMA device with a configurable circuit and sequencer that can perform transformations and searches by reading configuration information from the destination memory, applying specified operations to source data, and writing results back, allowing for transformations such as encryption and searching within the DMA transfer process, thereby reducing computational overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional DMA devices are used for memory transfer, then the DMA interface remains simple and legacy-compatible, but additional processing steps are required for transformations and searches, increasing computational expense
Solution Approach 1:
The patent merges the DMA transfer function with transformation and search capabilities into a single integrated operation. The circuit performs memory-to-memory transfer while simultaneously applying specified transformations (encryption, compression, etc.) and searching for keys, eliminating the need for separate processing steps and reducing overall computational expense.
Solution Approach 2:
The DMA device is designed with multi-functionality, capable of performing not only standard memory transfer but also various transformations (encryption, compression, formatting) and search operations. This universal approach allows a single device to handle multiple tasks that traditionally required separate processing units.
2Adaptability or versatility
If additional processing steps are added for transformations and searches, then transformation capability is improved, but computational overhead and processing time increase
Solution Approach 1:
The circuit enables continuous useful action by performing transformations and searches during the memory transfer operation itself, rather than as separate sequential steps. The transfer, transformation, and search operations occur simultaneously in a pipelined manner, maximizing resource utilization and eliminating idle processing time.
Solution Approach 2:
The transformation parameters and search keys are pre-configured in the circuit before the transfer operation begins. This preliminary configuration allows the circuit to execute transformations and searches immediately during data transfer without requiring additional setup time, thereby improving processing efficiency.
3Adaptability or versatility
If complex transformations are performed during DMA transfer, then processing capability is enhanced, but device complexity and configuration complexity increase
Solution Approach 1:
The circuit is designed to be self-configuring, automatically loading transformation parameters and search keys from predetermined memory locations. This self-service capability reduces the burden on programmers to manually configure complex transformation operations, as the circuit handles parameter loading and setup automatically based on pre-stored configuration data.
Solution Approach 2:
The patent uses pre-configured transformation parameter tables and key storage areas as intermediaries between the programmer and the complex transformation operations. Programmers simply specify high-level operations, and the circuit automatically retrieves and applies the appropriate parameters from these intermediary structures, simplifying the interface while maintaining complex processing capability.
Data Source
AI summary
A direct memory access (DMA) device provides for transforming source data as it is transferred to a destination memory space. The transformation can encompass a range of arithmetic logic unit (ALU) operations. The transformation can include discerning comparative matches in the source address space, such that matched-indice-reference-offsets are transferred to destination memory. A processor requesting the transfer can also configure the transformation to be completed by writing configuration data to memory and/or programming the DMA device. In transforming data as it is transferred, the DMA device can obviate time-consuming processing otherwise done after conventional DMA transfers.