HP 226824-001 - ProLiant - ML750 소개 매뉴얼 - 페이지 7

{카테고리_이름} HP 226824-001 - ProLiant - ML750에 대한 소개 매뉴얼을 온라인으로 검색하거나 PDF를 다운로드하세요. HP 226824-001 - ProLiant - ML750 10 페이지. Visualization and acceleration in hp proliant servers
HP 226824-001 - ProLiant - ML750에 대해서도 마찬가지입니다: 자주 묻는 질문 (4 페이지), 구현 매뉴얼 (35 페이지), 기술 백서 (12 페이지), 펌웨어 업데이트 (9 페이지), 개요 (20 페이지), 구현 매뉴얼 (26 페이지), 소개 매뉴얼 (22 페이지), 문제 해결 매뉴얼 (18 페이지), 구현 매뉴얼 (11 페이지), 설치 매뉴얼 (2 페이지), 구성 매뉴얼 (2 페이지), 소개 매뉴얼 (19 페이지), 업데이트 매뉴얼 (9 페이지), 업데이트 매뉴얼 (16 페이지), 소개 매뉴얼 (12 페이지), 기술 개요 (9 페이지)

Higher performance

The asynchronous serial links between the memory controller and AMB enable higher performance by

the memory subsystem. The two devices operate on independent clocks, so the transmitting device

embeds a bit clock with the data stream. The bit clock consists of a start bit that indicates the

beginning of a chunk of information, or data word, and a stop bit that indicates the end of the data

word. This eliminates skew between the clock and the data—an issue with parallel interfaces—and

allows very high data rates to be achieved.

As described previously, the outbound links exhibit one-half the peak theoretical throughput of the

inbound links. For FB-DIMMS with DDR2-533 DRAM devices, the outbound links have a theoretical

throughput of 2.1 GB/sec, or approximately one-half of 4.3-GB/s inbound links (Figure 6). Because

the serial links transport write and read data simultaneously, the total theoretical throughput of the FB-

DIMM channel is 6.4 GB/s (2.1 GB/s plus 4.3 GB/s). When using DDR2-667 DRAM, the total

theoretical throughput of the FB-DIMM channel is 8.0 GB/s (2.7 GB/s plus 5.3 GB/s)

Figure 6. Theoretical throughput of the FB-DIMM channel

FB-DIMM technology improves reliability through extended use of ECC and the introduction of a new

fail-over mechanism. FB-DIMMs extend the use of ECC to protect commands and address information

in the event of a soft error. This feature can help reduce instances where the soft error could disrupt

server performance and reliability. A second data protection mechanism called "bit lane fail-over"

correction identifies a bad data path and removes it from service.