Talk · International Conference on Artificial Intelligence, Computer, Data Sciences and Applications · February 5, 2026
Optimizing Block Size and Memory Usage for Image Processing on GPU Using CUDA
Mekhriddin Rakhimov, Mannon Ochilov, Shakhzod Javliev, Azizbek Khojamurotov
Abstract
In this article, the Canny algorithm was used to accelerate the computational processes by parallel processing in image processing. In this case, all stages of the Canny algorithm were implemented in the computer's graphics processor using Compute Unified Device Architecture technology, and in order to accelerate the computational process, the image data was optimally used from the graphics processor's global, shared memory, and texture memories. During the study, images of different sizes were tested in the global and shared memory of the graphics processor using the Compute Unified Device Architecture model to optimize the use of graphics processor memory, with block sizes of$8 \times 8,16 \times 16$and$32 \times 32$. For comparison, the program execution was measured on the central processor using the Open Computer Vision Library. Based on the results obtained, the$\mathbf{8 x 8}$blocks of the graphics processor were considered the most optimal for small-sized images, and the$\mathbf{3 2} \boldsymbol{\times} \mathbf{3 2}$blocks were considered the most optimal for large-sized images. The$16 \times 16$block configuration showed the highest efficiency for all images. Shared memory and$\mathbf{1 6 x 1 6}$block configuration showed$\sim 81$times faster performance compared to CPU on a$4096 \times 4096$image and$\boldsymbol{\sim} \mathbf{4 9 \%}$faster performance compared to global memory.
More by Mannon Ochilov
- PadSink-Track: A Calibrated Framework for Padding-Sink-Preserving, Alignment-Tracked Selective Cross-Attention Key–Value Cache Access in CPU-Based Whisper Speech RecognitionTechnologies · 2026
- A Cache-Aware Compression Framework: Deriving the Compression Target and Stopping Rule for Transformers on CPU-Class HardwareAI · 2026
- The Achievement of Speech Technologies in Enhancing the Rehabilitation of Hearing-Impaired PatientLecture notes in networks and systems · 2026
- Performance Evaluation of Parallel Programming Models for K-means Clustering on Parallel Computing SystemsInternational Siberian workshops and tutorials on electron devices and materials · 2026
