https://bugs.kde.org/show_bug.cgi?id=497565
--- Comment #12 from [email protected] --- Hello Maik, Thank you for the quick reply. I understand your point regarding the performance of a general purpose GPU which can sometimes be slow for DNN tasks. However, my report is specifically about the new Intel Core Ultra hardware. The key feature of this CPU is not the GPU, but the dedicated NPU (Neural Processing Unit), also known as Intel AI Boost. This is a specialized chip designed only for AI/DNN tasks (like face detection), and it is meant to be significantly faster and much more power-efficient than the CPU for this exact workload. The problem is that the OpenVINO backend (which this bug report requests) is the only way for OpenCV to access this new NPU. Without the OpenVINO backend compiled in, digiKam cannot use the primary AI acceleration feature of this entire new generation of mainstream processors. This is a different issue than a simple "GPU vs CPU" benchmark on older hardware. Could you please re-evaluate the importance of this bug in light of this new hardware? Thank you. -- You are receiving this mail because: You are watching all bug changes.
