[Image Processing Engineer Column] Misunderstandings about Acceleration
アポロ精工 アポロ精工 新横浜オフィス（旧ケーアイテクノロジー/旧市川ソフトラボラトリー）
We are developing and providing solutions through software and hardware, but it is not uncommon for there to be misunderstandings due to our mention of "acceleration using FPGA."
The belief that any processing implemented in software can be "accelerated" by developing a dedicated board is a significant misconception. In fact, the types of processing that can be accelerated using FPGA are limited.
It is suitable for processes that can execute the same computation in large quantities [simultaneously]. Convolution filters are a prime example of this. If you equip the necessary number of multiplication and addition circuits to perform cumulative operations [simultaneously], you can obtain results in one clock cycle, no matter how extensive the range.
Additionally, processes that allow for "pipelining" can transition from process A to process B in hardware, performing process A simultaneously when moving to process B, thus eliminating the time of the faster process. Even with a five-stage pipeline, results can be obtained in the time of just one stage.
By the way, when accessing large volumes of image data, GPUs with very wide data bus widths are also effective.
*News is distributed through our company newsletter.
Inquiry about this news
Contact Us Online
Distributors
