High Performance Implementation of Decoding Variable Length Codes on Media Processor
-
- ISHIGURO Masao
- Hitachi, Ltd. Systems Development Laboratory
-
- FUJIKAWA Yoshifumi
- Hitachi, Ltd. Systems Development Laboratory
-
- YAMAGUCHI Muneaki
- Hitachi, Ltd. Central Research Laboratory
-
- HIROI Kazushige
- Hitachi, Ltd. Systems Development Laboratory
-
- SUZUKI Norihiro
- Hitachi, Ltd. Central Research Laboratory
-
- KAWAGUCHI Atsuo
- Hitachi, Ltd. Systems Development Laboratory
Bibliographic Information
- Other Title
-
- メディアプロセッサにおける可変長復号処理の高速化(マルチメディア通信とQoS, VoIP, ストリーミング, 一般)
- メディアプロセッサにおける可変長復号処理の高速化
- メディアプロセッサ ニ オケル カヘンチョウフクゴウ ショリ ノ コウソクカ
Search this article
Abstract
SIMD-type instructions and/or VLIW architecture are well-known features of high-performance processors. Although these features are effective for data-parallel algorithms, they have (almost) no effect for sequential algorithms such as decoding variable length codes. In order to address this issue, the media processor MAPCA has a co-processor for processing variable length code in addition to its 136-bit VLIW processor. Our prototype MPEG-4 decoder runs about 3.1 times faster than that using only the VLIW processor on MAPCA.
Journal
-
- ITE Technical Report
-
ITE Technical Report 26.48 (0), 19-24, 2002
The Institute of Image Information and Television Engineers
- Tweet
Details 詳細情報について
-
- CRID
- 1390282679502568320
-
- NII Article ID
- 110003671556
-
- NII Book ID
- AN1059086X
-
- ISSN
- 24241970
- 13426893
-
- NDL BIB ID
- 6295768
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- NDL
- CiNii Articles
-
- Abstract License Flag
- Disallowed