HP 226824-001 - ProLiant - ML750 Einführung Handbuch - Seite 2

Blättern Sie online oder laden Sie pdf Einführung Handbuch für Desktop HP 226824-001 - ProLiant - ML750 herunter. HP 226824-001 - ProLiant - ML750 10 Seiten. Visualization and acceleration in hp proliant servers
Auch für HP 226824-001 - ProLiant - ML750: Häufig gestellte Fragen (4 seiten), Handbuch zur Umsetzung (35 seiten), Technisches Weißbuch (12 seiten), Firmware-Aktualisierung (9 seiten), Übersicht (20 seiten), Handbuch zur Umsetzung (26 seiten), Einführung Handbuch (22 seiten), Handbuch zur Fehlersuche (18 seiten), Handbuch zur Umsetzung (11 seiten), Installationshandbuch (2 seiten), Konfigurationshandbuch (2 seiten), Einführung Handbuch (19 seiten), Handbuch aktualisieren (9 seiten), Handbuch aktualisieren (16 seiten), Einführung Handbuch (12 seiten), Technologie-Brief (9 seiten)

Seite drucken
HP 226824-001 - ProLiant - ML750 Einführung Handbuch

Abstract

This paper describes the features, benefits, and challenges of Fully-Buffered dual inline memory
module (FB-DIMM) technology. It also provides rules for populating FB-DIMM slots to achieve
maximum performance in HP ProLiant servers.

Introduction

HP ProLiant servers provide balanced system architectures that deliver peak performance per watt of
power. A balanced system architecture is one in which the three main server subsystems—processing,
memory, and I/O—interact efficiently to maximize CPU performance. The introduction of dual-core
processor technology challenged HP engineers to improve the performance of the memory and I/O
subsystems to maintain system balance. Engineers overcame the performance bottleneck of the
parallel I/O bus by migrating to a high-speed serial interface with technologies such as Serial-
Attached SCSI, Serial ATA, and PCI Express. Likewise, some memory subsystems are migrating to
high-speed serial FB-DIMM technology to deliver scalable bandwidth and memory capacity that is not
possible with traditional DIMM technologies.
The paper describes the barriers that limit memory capacity and performance of servers that use
traditional DIMM technologies. This paper also describes the operation of FB-DIMMs and summarizes
the benefits and challenges of using FB-DIMM technology.

Performance barriers for traditional DIMM

Traditional DIMM architectures use a stub-bus topology with parallel branches (stubs) that connect to
a shared memory bus (Figure 1). The memory bus consists of the command/address (C/A) bus and
the data bus. The C/A bus consists of 21 data traces that transport command and address signals to
the DIMMs. The data bus consists of 72 traces, each carrying one bit at a time (a total of 64 data bits
and 8 ECC
bits). Each DIMM connects to the data bus using a set of pin connectors
1
electrical signals from the memory controller to reach the DIMM bus-pin connections at the same time,
all the traces have to be the same length. This often results in circuitous traces on the motherboard
between the memory controller and memory slots. Both the latency (delay) resulting from complex
routing of traces and signal degradation at the bus-pin connections cause the error rate to increase as
the bus speed increases.
Figure 1. Stub-bus topology.
Error correcting code
1
Typical SDRAM DIMMs have a total of 168 pins, DDR DIMMs have 184 pins, and DDR2 DIMMs have 240 pins.
2
An impedance discontinuity is created at each stub-bus connection.
. In order for the
2
2