학술논문
Intel Wireless MMX 기술을 이용한 H.264 인코더 구현 / Implementation of H.264 encoder using Intel Wireless MMX technology
Document Type
Dissertation/ Thesis
Author
Source
Subject
Language
Korean
Abstract
이전의 표준과 비교해보면 H.264의 코딩의 능률과 코딩의 성능은 높다. 그러나 그것은 많은 메모리 엑세스와 많은 수의 계산양이 필요하다. Intel Wireless MMX™ 기술은 효과적인 파워 방법에 모바일 멀티미디어 가속과 어플리케이션 프로세싱을 하기 위하여 디자인 되었다. SIMD의 효율적인 구조는 엔코딩 프로세스의 병렬처리 창출에 필요한 명령어 셋을 제공할 수 있다. 병렬처리의 효과는 컴퓨터 처리 시간의 축소와 뿐만 아니라 메모리 엑세스 시간의 축소도 명백하다.본 논문은 Marvel반도체사에서 만들어진 강력한 임베디스 프로세서와 같은 PXA320-PRO 하드웨어 플렛폼을위한 H.264/AVC Baseline Profile 엔코더 최적화 방법에 대하여 소개한다. 중점은 "X264"의 소스 프로그램에 채택된 시스템의 최적화 그리고 알고리듬 최적화 또한 명령어 최적화이다. 목표는 SIMD 명령어 세트와 메모리 관리의 최적화 사용에 의한 Intel Wireless MMX™ 기술에 H.264 구현의 향상이다. H.264 표준을 위한 코드 최적화는 모바일 Intel Wireless MMX™ 가 지원되는 플렛폼에 구현 되었다. 최적화된 엔코더의 코딩 성능은 작은 손실과 함께 향상되었고 그리고 QCIF 해상도 결과는 실시간으로 엔코딩 되는 것을 결과는 보여 진다.
Compared with the previous standards, the coding efficiency and coding performance of H.264/AVC is improved at the cost of higher computational complexity and large memory access. Intel Wireless MMX™ technology has been designed to accelerate mobile multimedia and applications processing in a power efficient manner. The availability of SIMD architecture can provide the necessary instruction set to introduce the parallelism in the encoding process. The effect of parallelism is evident not only in reduction of computational time, but also in reduction in memory access time reduction.This paper introduces the method to optimize the H.264/AVC Baseline Profile encoder for PXA320-PRO hardware platform which has a powerful embedded processor PXA320 produced by Marvel. The emphasis is on the system optimization, algorithm optimization and instruction optimization which has been adopted in the "x264" source program. The target plan is to improve the H.264 implementation on Wireless MMX™ technology by using SIMD instruction set and optimizing the memory management. The optimized codes for the H.264 standard is implemented on mobile Wireless MMX™ supported platform. Results show that the coding performance of the optimized encoder is improved with a little loss and video sequences with QCIF resolution can be encoded in real time.
Compared with the previous standards, the coding efficiency and coding performance of H.264/AVC is improved at the cost of higher computational complexity and large memory access. Intel Wireless MMX™ technology has been designed to accelerate mobile multimedia and applications processing in a power efficient manner. The availability of SIMD architecture can provide the necessary instruction set to introduce the parallelism in the encoding process. The effect of parallelism is evident not only in reduction of computational time, but also in reduction in memory access time reduction.This paper introduces the method to optimize the H.264/AVC Baseline Profile encoder for PXA320-PRO hardware platform which has a powerful embedded processor PXA320 produced by Marvel. The emphasis is on the system optimization, algorithm optimization and instruction optimization which has been adopted in the "x264" source program. The target plan is to improve the H.264 implementation on Wireless MMX™ technology by using SIMD instruction set and optimizing the memory management. The optimized codes for the H.264 standard is implemented on mobile Wireless MMX™ supported platform. Results show that the coding performance of the optimized encoder is improved with a little loss and video sequences with QCIF resolution can be encoded in real time.