HP ProLiant AMD-based 300–series G6 and G7 servers

HP ProLiant AMD-based 300–series G7
servers
Technology brief
Abstract.............................................................................................................................................. 2
Introduction......................................................................................................................................... 2
HP servers and balanced architecture .................................................................................................... 2
Processor technology............................................................................................................................ 3
Direct Connect architecture 2.0 ......................................................................................................... 4
HyperTransport technology ............................................................................................................... 4
HT Assist ......................................................................................................................................... 5
Memory technologies ........................................................................................................................... 5
DDR3 DIMM choices ........................................................................................................................ 5
Memory management technologies .................................................................................................... 6
X8 error correction ........................................................................................................................... 7
I/O technologies ................................................................................................................................. 7
PCI Express technology..................................................................................................................... 8
HP Smart Array and SAS/SATA technology ........................................................................................ 8
Power and thermal technologies .......................................................................................................... 11
Efficient power delivery................................................................................................................... 11
Improved thermal sensors and fan control ......................................................................................... 12
HP Dynamic Power Capping ........................................................................................................... 13
Managing processor technologies.................................................................................................... 14
Systems management and monitoring .................................................................................................. 16
HP Integrated Lights-Out 3............................................................................................................... 16
HP Insight Control Environment ........................................................................................................ 16
HP Insight Dynamics ....................................................................................................................... 16
Security ............................................................................................................................................ 17
OS support ....................................................................................................................................... 17
Summary .......................................................................................................................................... 17
For more information.......................................................................................................................... 18
Call to action .................................................................................................................................... 19
Abstract
This technology brief describes the key technologies implemented in HP ProLiant 300-series G7
servers based on AMD™ processors. As of this writing, the AMD-based 300--series G7 server
platforms are limited to the ProLiant DL385. For detailed information about this server, refer to the
QuickSpec link listed at the end of this technology brief.
Introduction
The HP ProLiant DL385 G7 server includes these key technologies:
• AMD Opteron™ eight- and twelve-core 6100 Series processors
• Thermal sensors incorporated throughout the ProLiant 300-series G7 servers
• BIOS controlled memory and processor management capabilities
• Integrated Lights-Out 3 (iLO 3) remote server management and control
• I/O technologies such as PCI Express generation 2 (PCIe 2.0) and Smart Array controllers that
incorporate common form factor components
• Flash-backed write cache for Smart Array controllers
• Common-slot power supplies in multiple sizes to provide the required amount of power and improve
power efficiency
HP servers and balanced architecture
HP designs cost-competitive, power-efficient servers that use a balanced architecture to address
performance requirements and provide value.
HP servers achieve a balanced architecture through superior engineering of fundamental elements
such as mechanical infrastructure, power, cooling, processor, memory, IO devices, storage, boot,
networking, and interconnecting components. A balanced architecture includes the following aspects:
• Compute capability, processor core count, cache size per processor, and processor socket count
• Low-latency processor-to-memory bandwidth commensurate with core count
• Memory footprint and capacity that maximizes bandwidth and capacity with power efficiency and
performance without compromising quality or reliability
• Application-appropriate IO devices
• Closely-coupled and balanced processor-to-memory and processor-to-I/O ratios
• Mechanical design that ensures optimum levels of cooling, stability, and serviceability through
space-efficient, modular partitioning across the server
By designing a balanced architecture, HP ensures that all subsystems can be used effectively under a
broad range of applications and workloads. For example, increasing memory capacity
asymmetrically will not increase performance as effectively as distributing the same amount of
memory across processors and IO devices. Inefficient memory distribution yields diminishing returns
on power consumption and cost. A virtual machine (VM), for example, benefits from memory closely
coupled to the processor responsible for that VM. Furthermore, a server needs to have appropriate
levels of I/O bandwidth and CPU capabilities to ensure that memory can be used effectively by every
VM.
2
Processor technology
The HP AMD-based 300-series G7 servers use the AMD Opteron 6100 series eight- and twelve-core
processors allowing up to 24 cores in a two-processor platform. These processors are based on
AMD's 45 nanometer process and have a core standard wattage of 75W Average CPU Power. In
addition, these processors use Direct Connect™ architecture 2.0, HyperTransport™ 3.0, HT Assist™,
and 8x ECC error correction. The 6100 processors fit into the G34 socket infrastructure and feature
DDR3 memory, four 64 bit memory channels per processor, the AMD SR5690/SP5100 chipset, and
quad HyperTransport (HT) links as shown in Figure 1. The processors operate at speeds of up to 2.4
GHz with eight cores and 2.3 GHz with twelve cores. They access 512 KB of L2 cache memory per
core and share a 12 MB L3 cache. The AMD Opteron 6100 series processor includes performanceoptimized power and thermal controls. See the Power and thermal technologies section for more
information.
Figure 1. DL385 G7 server architecture using the AMD 6100 series processor and chipset
GFX_2 15:8
VRD 2
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
10.4GB/s
HT L2
Proc 2
GFX_2 7:0
SR5690
PCIE Gen II x8
Riser 1
PCIE Gen II x4
PCIE Gen II x4
10 GB/s
10 GB/s
GFX_1 15:8
GFX_1 7:0
PCIE Gen II x8
Riser 2
PCIE Gen II x4
PCIE Gen II x4
10 GB/s
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
Memory
PCIE SB 3:0
1 GB/s
5.2GB/s
5.2GB/s
L1 L0 L3A L3B
Memory
GPP 9:6
4 GB/s
L1 L3A L3B
10.4GB/s
10.4GB/s
L2
Proc 1
VRD 1
SP5100
GPP 1:0
1 GB/s
PMC8011
5709
GPP 3:2
1 GB/s
5709
GPP 4
0.25 GB/s
iLO3
32-bit PCI
Video
USB 6 Ports
USB
SATA Optical Port
FL FH Slot 1
HL FH Slot 2
HL FH Slot 3
FL FH Slot 4
LP Slot 5
LP Slot 6
10 GB/s
1Gb
1Gb
1Gb
1Gb
iLO 10/100Mb port
6 USB ports
CD/DVD
NOTE:
In this AMD Opteron 6100-series architecture, the SR5690 (sometimes referred to as
the Northbridge chip) is responsible for I/O and PCI devices, while the SR5100
(sometimes referred to as the Southbridge chip) is responsible for USB, CD/DVD, and
video functions
3
Direct Connect architecture 2.0
Unlike front-side bus architecture, Direct Connect integrates the memory controller into the processor
and directly connects CPUs to the I/O subsystem and other processors (Figure 2). Direct Connect
architecture uses direct HyperTransport links between CPUs, between CPU and I/O, and between
CPU and memory. Direct Connect architecture currently scales up to 12 cores to provide superior
memory and I/O capability, near native virtualization performance, and a range of power bands 1
that place a priority on low power consumption.
HT3
L1
10.4GB/s
L1
HT3
5.2GB/s
10.4GB/s
L2
L3B
HT3 L2
5.2GB/s
L3B
72 Bit
12 MB L2 cache
L3A
72 Bit
HT3
512 KB shared cache per core
L3A
DDR3
L2
10.4GB/s
L0
8 or 12 Cores
72 Bit
L1
AMD Opteron 6100-series
processor
72 Bit
HT3
Figure 2. Block diagram of Direct Connect 2.0 architecture in the AMD 6100 series processors
HyperTransport technology
HyperTransport is a point-to-point interconnect that is designed to connect the processors directly and
to connect each processor to its dedicated memory banks, as well as to other I/O chipsets. 2
Compared to a shared, parallel front-side bus, the advantages of HyperTransport include no
overhead for bus arbitration and easier signal integrity maintenance, which results in a scalable, highbandwidth architecture.
Each16-bit (2-byte) HyperTransport link is double-pumped to perform two data transfers per clock
cycle. From HyperTransport 1.0 (HT1) in 2001 to HyperTransport 3.1 in 2008, the maximum clock
1
2
Power bands refer to a new metric developed by AMD to reflect power consumed by the processor and its integrated memory
controller during peak workloads. This metric is based on AMD’s measurement of Average CPU Power (ACP). For more
information on ACP, see the whitepaper at www.amd.com/usen/assets/content_type/white_papers_and_tech_docs/43761C_ACP_WP.pdf
HyperTransport Technology was invented at AMD with contributions from industry partners and is managed and licensed by
the HyperTransport Technology Consortium, a Texas non-profit corporation.
4
speed and transfer rate increased from 800 MHz (1.6 MT/s 3 ) to a maximum of 3.2 GHz (6.4 GT/s)
in each direction.
The G34 socket infrastructure features four, 16-bit HT3 links operating at up to 6.4 GT/s per link. Of
the four HT3 links, three are shared processor to processor links and one link on processor one is
used for I/O communication.
HT Assist
HT Assist reduces cache coherence traffic on the HT links. By tracking where data is stored in cache
and guiding the processor directly to the cache of other processors, HT Assist reduces cache probe
traffic between processors, especially in 4- socket servers. Therefore, HT Assist results in faster queries
that can increase performance for cache-sensitive applications such as database, virtualization, and
compute-intensive applications.
Memory technologies
In the AMD Opteron 6100-series architecture, the memory controller is integrated into the processor
chip to optimize memory performance and bandwidth per CPU. The memory controller reduces
latency inherent in front side bus architectures by eliminating the bus contention between memory and
I/O cycles. A server’s overall memory bandwidth increases as processors are added to a
configuration, unlike legacy designs that scale poorly because access to main memory is limited by
external Northbridge chips. HP designed specific BIOS and RBSU functions to manage memory
configurations, letting the customer optimize configurations for maximum performance while reducing
power consumption and cooling requirements. These management options also include memory
protection and latency reduction. Because of the increased reliability of DDR3 on-DIMM thermal
sensors, HP incorporates DIMM thermal data into the algorithms controlling thermal and power states
within the server.
DDR3 DIMM choices
HP ProLiant 300-series G7 servers with Opteron 6100 series processors support double data rate
(DDR3) DIMMs. DDR3 has several key enhancements including an 8-bit prefetch buffer for storing
data before it is requested. By comparison, DDR-2 has a 4-bit buffer. For DDR3, the data signal rate
can increase to 1333 Megatransfers per second (MT/s). While this is commonly referred to as having
a speed of 1333 MHz, the maximum clock speed for the DIMMs is actually 667 MHz and the signal
is double-pumped to achieve the 1333 MT/s data rate. DDR3-1333 DIMMs can operate at clock
speeds of 667 MHz, 533 MHz, and 400 MHz with corresponding data rates of 1333, 1066, and
800 MT/s.
HP DDR3 DIMM modules incorporate an integrated thermal sensor that signals the processor to
throttle memory traffic to the DIMM if its temperature exceeds a programmable critical trip point.
Using the data from these thermal sensors, ProLiant AMD-based 300-series G7 servers can reduce fan
speed when memory is idle, which reduces power consumption. The BIOS in ProLiant G7 servers
verifies the presence of the thermal DIMM sensor during POST. Some third-party DIMMs may not
include this thermal sensor. If it is absent, a POST message will warn that the DIMM does not have a
thermal sensor, and the fans will be forced to run at higher speeds (requiring more power).
DDR3 is available as both Unbuffered Dual In-line Memory Modules (UDIMMs) and Registered
(buffered) Dual In-line Memory Modules (RDIMMs). Both RDIMMs and UDIMMs support error
correcting code (ECC).
3
MT/s, or megatransfers per second, equals the speed of the link in millions of cycles per second times the number of transfers
per cycle.
5
There are three types of standard voltage DDR3 available for ProLiant 300 series G7 servers:
• PC3-8500R- DDR3 (RDIMM, ECC compliant)—1333 or 1066 MT/s data rate depending on
memory configuration.
• PC3-10600R DDR3 (RDIMM, ECC compliant)—1333 or 1066 MT/s data rate depending on
memory configuration.
• PC3-10600E DDR3 (UDIMM, ECC compliant)—1333 MT/s data rate
Low-voltage DDR3
Customers can take advantage of the HP low voltage (LV) DDR3 memory option. LV memory can
operate at 1.35 V, reducing power and cooling requirements. Some LV DIMM configurations can
affect performance. The difference occurs in 2 DIMM per channel (DPC) configurations with dual-rank
DIMMs. In these configurations, the memory bus runs at 1.5 V with 2 DPC at 1333 MT/s, or 1.35 V
with 2 DPC at 1066 MT/s. This results in a 20% reduction in bandwidth for the lower voltage. In all
other HP LV configurations, the LV data rate is the standard 1.5V data rate. The BIOS determines the
operating voltage and data rate capability from the DIMMs and the DIMM population from the
system. The BIOS then sets the data rate based on that information.
NOTE:
Although the bandwidth reduction from DDR3-1333 to DDR3-1066 is 20%, the
measured reduction in throughput is 10%
Single rank DIMMs configured at 1, 2, and 3 DPC; and dual rank DIMMs configured at 1 and 3 DPC
run at the 1333 MT/s data rate at both voltages.
ProLiant G7 server models with the AMD 6100-Series processor support LV memory:
• PC3L-10600R—1333 1066 or 800 MT/s data rate, depending on memory configuration.
DIMM configuration guidelines
ProLiant AMD-based 300-series G7 servers support 3 DIMMs per channel up to 24 DDR3 memory
DIMMs. DDR3 DIMM speeds will vary depending on number of DIMMs per channel. Consult the
server QuickSpecs to determine DIMM speeds for given configurations. Administrators can configure
Opteron-based ProLiant 300-series G7 servers using either RDIMMs or UDIMMs, but RDIMM and
UDIMM memory cannot be mixed within a single server. Low voltage DIMMs can provide up to 10%
DIMM power savings. Low and standard voltage DIMMs are compatible; systems automatically adjust
voltage based on DIMMs installed.
NOTE:
DDR3 DIMM speeds will vary depending on number of DIMMs per channel. Consult
the server QuickSpec to determine DIMM speeds for given configurations.
For detailed memory configuration rules and guidelines, please use the Online DDR3 Memory
Configuration Tool at www.hp.com/go/ddr3memory-configurator.
Memory management technologies
Memory interleaving on the AMD Opteron 6100-series processors can occur between the processor
memory banks, memory channels, and between processor nodes in a multiprocessor system. ProLiant
6
AMD-based 300-series G7 servers support all three memory interleaving technologies. Even though
these technologies are independent of each other, they can operate simultaneously.
Memory bank interleaving
With memory bank interleaving engaged, data is routed alternately to memory banks through the
common memory channel connecting the DIMM banks and the integrated memory controller.
However, memory bank interleaving does increase the probability that more DIMMs need to be kept
in an active state (requiring more power) since the memory controller alternates between memory
banks and therefore between DIMMs.
The processor node memory controller automatically enables memory bank interleaving under the
following conditions:
• Two single-rank DIMMs per channel result in two way-bank interleaving
• Two dual-rank DIMMs per channel result in four way-bank interleaving
• Two quad-rank DIMMs per channel results in eight-way bank interleaving
• Two dual-rank DIMMs and one quad-rank DIMM results in eight-way bank interleaving (in servers
using three DIMMs per channel)
Memory Channel Interleaving
With memory channel interleaving, the processor memory controller routes data alternately through
the two available memory channels. The result is that when the memory controller needs to access a
block of logically contiguous memory, the requests are distributed more evenly across the two
channels rather than potentially stacking up in the request queue of a single channel. This alternate
routing decreases memory access latency and increases performance. As with memory bank
interleaving, memory channel interleaving increases the probability that more DIMMs need to be kept
in an active state.
Memory channel interleaving is always active on the AMD 6100-series processor.
Memory node interleaving
With node interleaving, memory can be interleaved across any subset of nodes in the multiprocessor
system. Node interleaving breaks memory into 4 KB addressable entities. Addressing starts with
address 0 on node 0 and assigns sequential addresses through address 4095 to node 0, addresses
4096 through 8191 to node 1, addresses 8192 through 12287 to node 2, and addresses 12888
through 16383 to node 3. Address 16384 is assigned to node 0, and the process continues until all
memory has been assigned in this fashion. An application that uses a common allocation thread will
benefit from node interleaving.
Memory node interleaving is disabled by default. Administrators can activate node interleaving using
the RBSU. Node interleaving can only be configured if the memory footprint for both processors
identical.
X8 error correction
In AMD Opteron 6100-series processors, the memory controller supports error correction circuitry
(ECC) for both x4 and x8 DIMMs.
I/O technologies
ProLiant 300-series G7 servers incorporate PCI Express, Serial-Attached SCSI (SAS), and Serial ATA
(SATA) I/O technologies. This server architecture lets administrators add PCI Express-compliant
7
expansion cards to the system. SAS is a serial communication protocol for direct-attached storage
devices such as SAS and SATA hard drives.
PCI Express technology
All ProLiant G7 servers support the PCIe 2.0 specification. PCIe 2.0 has a per-lane signaling rate of
5 Gb/s―double the per-lane signaling rate of PCIe 1.0. PCIe 2.0 is completely backward compatible
with PCIe 1.0. A PCIe 2.0 device can be used in a PCIe 1.0 slot and a PCIe 1.0 device can be used
in a PCIe 2.0 slot. Table 1 shows the level of interoperability between PCIe cards and PCIe slots.
Table 1. PCIe device interoperability
PCIe
device type
x4 Connector
x4 Link
x8 Connector
x4 Link
x8 Connector
x8 Link
x16 Connector
x8 Link
x16 Connector
x16 Link
x4 card
x4 operation
x4 operation
x4 operation
x4 operation
x4 operation
x8 card
Not allowed
x4 operation
x8 operation
x8 operation
x8 operation
x16 card
Not allowed
Not allowed
Not allowed
x8 operation
x16 operation
HP Smart Array and SAS/SATA technology
The latest serial PCIe 2.0-capable Smart Array controllers use Serial Attached SCSI (SAS) technology,
a point-to-point architecture in which each device connects directly to a SAS port rather than sharing
a common bus as with parallel SCSI devices. Point-to-point links increase data throughput and
improve the ability to locate and fix disk failures. More importantly, SAS architecture solves the
parallel SCSI problems of clock skew and signal degradation at higher signaling rates. 4
The latest Smart Array controllers are compatible with Serial Advanced Technology Attachment
(SATA) technology and include the following features to enhance performance and maintain data
availability and reliability:
• SAS and SATA compatibility — SAS-2 compliance lets administrators deploy and manage both
SAS arrays and SATA arrays. Smart Array configuration utilities help administrators configure
arrays correctly so that data remains available and reliable.
• SAS wide port operations — Wide ports contain four single lane (1x) SAS connectors and the
cabling bundles all four lanes together. SAS wide ports enhance performance by balancing SAS
traffic across the links. In addition, wide ports provide redundancy by tolerating up to three physical
link failures while maintaining the ability to communicate with the disk drives. The tolerance for link
failures is possible because wide port connections are established from Phy 5 to Phy, and multiple,
simultaneous connections to different destinations are possible. The most common use of these wide
links is to a JBOD or to an internal server expander connecting to large numbers of drives. No
special configuration is required for this functionality.
• SAS expanders — Low-cost, high-speed switches called expanders can combine multiple single
links to create wide ports and increase available bandwidth. SAS expander devices also offer
higher system performance by expanding the number of hard drives that can be attached to an
HP Smart Array controller. SAS expanders are an aggregation point for large numbers of drives or
servers providing a common connection. By cascading expanders, administrators can chain
multiple storage boxes together.
4
For more information about SAS technology, refer to the HP technology brief titled “Serial Attached SCSI storage technology”
available at http://h20000.www2.hp.com/bc/docs/support/SupportManual/c01613420/c01613420.pdf .
5
The mechanism that contains a transceiver which electrically interfaces to a physical link. Phy is a common abbreviation for
the physical layer of the OSI model.
8
For more information on the HP SAS Expander Card, go to
http://h18004.www1.hp.com/products/servers/proliantstorage/arraycontrollers/sasexpander/index.html.
SAS-2
SAS-2 and PCIe 2.0 are among the technologies responsible for a significant increase in performance
over past generations of Smart Array controllers. The second-generation SAS (SAS-2) link speed 6 of
6.0 Gb/s is double the SAS-1 transfer rate. Operation at SAS-2 link speeds requires SAS-2 compliant
hard drives. SAS-2 eliminates the distinction between fanout and edge expanders by replacing them
with self-configuring expanders. SAS-2 enables zoning for enhanced resource deployment, flexibility,
security, and data traffic management. SAS-2 is also backward compatible with SAS-1.
Beginning with HP product releases in the first quarter of 2009, Smart Array controllers are SAS-2
capable. In fully supported controllers, 6-Gb/s SAS technology allows Smart Array controllers to
deliver peak data bandwidth up to 600 MB/s per physical link in each direction. SAS devices are
capable of sending and receiving data simultaneously across each physical link (full duplex mode).
When running full duplex, 6-Gb/s SAS technology can deliver peak data bandwidth up to 1200
MB/s.
The SAS-2 specification is compatible with both Serial SCSI and Serial ATA protocols for
communicating commands to SAS and SATA devices. SAS-2 compliant controllers are fully
compatible with 1.5-Gb/s and 3.0-Gb/s SATA technology.
For the most up-to-date listing of HP Smart Array controllers that support the SAS-2 specification, see
the Smart Array controller matrix: www.hp.com/products/smartarray.
HP Smart Array controllers based on PCIe 2.0
The Smart Array PCIe 2.0-based controllers are modular solutions with a common form factor,
hardware, and firmware. Any of the ProLiant 300-series G7 servers can use PCIe 2.0-based
controllers. As a standard entry level RAID, HP designed the Smart Array 410i with a unique Zero
Memory RAID (ZMR) capability. Administrators can choose the cache size and whether to include the
battery backed write cache (BBWC) or the flash-backed write-cache (FBWC). These options allow
users to upgrade from ZMR to 512 MB BBWC or 1GB FBWC.
Battery backed write cache
The BBWC system continues to be an option for capacity expansion (adding one or more physical
disks to an existing array). The Smart Array controller recalculates parity and balances the data
across all the disks. During the expansion, the BBWC preserves data and logical structures on the
array. The HP 650 mAh P-Series battery can power the cache for up to 48 hours before recharging
becomes necessary.
Flash-backed write cache
HP introduced the flash-backed write-cache (FBWC) system in the fourth quarter of 2009. The FBWC
uses NAND 7 flash devices to retain cache data and super-capacitors (Super-caps) instead of batteries
to provide power during a power loss. The FBWC offers significant advantages over the HP Batterybacked write-cache (BBWC) system. Since the FBWC writes the contents of memory to flash devices,
there is no longer a 48-hour battery life limitation and the data will be posted to the disk drive on the
next power up.
6
7
The Serial Attached SCSI - 2 (SAS-2 or SAS 2.0) is a draft standard, and is the product of the Technical Committee 10t
Organization. SAS 2.0 is second generation of SAS and is based upon SAS - 1.1. The SAS-2 specification is available from
the 10t website, http://www.t10.org.
Non-volatile semiconductor memory that can be electronically erased and reprogrammed. No power is needed to maintain
data stored in the chip
9
The FBWC DDR2 mini-DIMM cache module is specifically designed for the present generation of
PCIe2.0, SAS-based Smart Array controllers based on the PMC PM8011 SAS SRC 8x6G RAID on a
chip (RoC). The primary FBWC components consist of the cache module, Super-caps with integrated
charger, and RoC located on the system board.
At the time of this writing, the FBWC cache is supported on the Smart Array P410, P410i, P411,
P212, P812, and P712m.
For more information on the flash-backed write cache, see the “HP Smart Array Controller technology
brief” at
http://h20000.www2.hp.com/bc/docs/support/SupportManual/c00687518/c00687518.pdf
Zero Memory RAID
Using Zero Memory RAID (ZMR), administrators can create a RAID 0-1 configuration without
additional memory. Smart Array P410, P411, and P212 controllers include ZMR. The P212 controller
does not include ZMR on the external connector. ZMR supports up to eight drives in Zero Memory
Mode, or seven drives and one tape drive. ZMR does not include caching; however, all systems can
be upgraded to a BBWC or FBWC memory module that can significantly increase performance.
NOTE:
Smart Array Advanced Pack is not available on Zero Memory configurations.
Smart Array Advanced Pack
HP Smart Array Advanced Pack (SAAP) firmware provides advanced functionality within Smart Array
controllers. This firmware further enhances performance, reliability, and data availability. The Smart
Array controller hardware firmware stack supports SAAP. It can be enabled on the P212, P410,
P410i, and P411 controllers.
SAAP requires a license key for activation. After activation, administrators can use several
capabilities:
• RAID 6 with Advanced Data Guarding (ADG) protects against failure of any two drives. It requires
a minimum of four drives, but only two will be available for data. ADG can tolerate multiple
simultaneous drive failures without downtime or data loss and is ideal for applications requiring
large logical volumes, because it can safely protect a single volume of up to 56 disk drives. RAID
ADG also offers lower implementation costs and greater usable capacity per U than RAID 1.
• RAID 60 allows administrators to split the RAID storage across multiple external boxes. It requires a
minimum of eight drives, but only four will be available for data.
• Advanced Capacity Expansion (ACE) automates higher capacity migration using capacity
transformation to remove logical drives by shrinking and then expanding them online. Standard
drive migration and expansion remain unchanged.
• Mirror Splitting and Recombining in Offline Mode breaks a RAID 1 configuration into two RAID 0
configurations. This is similar to a scaled down rollback functionality that requires two disk drives.
• Drive Erase completely erases physical disks or logical volumes. This capability is useful when
decommissioning, redeploying, or returning hard drives.
• Video On Demand Performance Optimization decreases latency and improves video streaming.
More information about SAAP is available at www.hp.com/go/SAAP.
10
NOTE:
At a minimum, a 256 MB cache and battery kit is required to enable the
SAAP license key. SAAP is not available on Zero Memory Configurations.
Solid state drives
HP has introduced the second generation of solid state drives (SSD) for ProLiant servers. These solid
state drives are 3Gb/s SATA interface in both 60GB and 120GB capacities. The product, based on
NAND Single Level Cell flash technology, are implemented as SFF and LFF hot plug devices on the HP
universal drive carrier for general use across the ProLiant portfolio. These drives deliver higher
performance, lower latency, and low power solutions when compared with traditional rotating media.
You can use the HP second generation SSDs with the present generation Smart Array controllers
based on the PM8011 SRC MIPS processor on select ProLiant G6 and G7 servers. See the server
QuickSpecs to confirm that SSDs are supported.
For more information on HP second generation SSDs, download the “Drive technology overview”
technology brief at
http://h20000.www2.hp.com/bc/docs/support/SupportManual/c01071496/c01071496.pdf
Power and thermal technologies
HP engineers have developed a robust set of power and thermal technologies and components to
manage power within ProLiant 300-series G7 servers. The following technologies improve power
efficiency throughout the power delivery chain:
• Efficient power delivery
• Thermal sensors and fan control
• Dynamic Power Capping
• Processor Management technologies
Administrators can disable certain components and capabilities within ProLiant 300-series G7 servers
or reduce capabilities to bring the components to a lower power state.
Efficient power delivery
The ProLiant DL385 G7 server uses HP common slot power supplies. To see the power supply options
offered with other HP servers, consult the server QuickSpecs or go to the HP ProLiant Server
Compatibility Guide at www.hpproliantoptions.com/intro/.
Common Slot power supplies
The HP Common Slot power strategy provides power supply commonality across supported ProLiant
G7 servers. Three different sized common slot power supplies are available so customers can choose
the most effective power supply to match their power needs. “Right sizing” power supplies lets
customers more closely match the power supply to the server power requirements in specific
environments, significantly reducing wasted power. The HP Common Slot power strategy has also
reduced the number of power supply designs, which in turn reduces the number of spares the
customer must keep in the data center.
Power supply efficiency relates to the level of effective transfer and delivery of power through the
power chain. Table 2 shows that HP power supplies have achieved efficiency ratings of up to 94%,
meeting the Climate Savers Platinum requirements.
11
Table 2. HP power supply efficiency and Climate Savers rating
Power supply
Efficiency
Rating
460W AC
up to 92% efficiency
Climate Savers Gold
750W AC
up to 92% efficiency
Climate Savers Gold
1200W AC
up to 94% efficiency
Climate Savers Platinum
1200W 48VDC
up to 90% efficiency
Climate Savers Silver
Use the HP Power Advisor to help determine which power supplies will best meet your needs:
www.hp.com/go/hppoweradvisor.
Redundant power operation
With ProLiant G7 servers, customers can select a power supply operation mode for redundant power
systems. Through the RBSU, administrators can select “High Efficiency” mode or “Balanced” mode. In
Balanced mode, both power supplies provide power equally. This mode ensures full redundancy but
can result in higher power consumption when power supplies are operating with reduced loads and
lower power efficiency. In High Efficiency mode, the system will only use one power supply until
system load exceeds a certain threshold. The second power supply stays online maintaining
redundancy but does not supply power until needed. Either selection still provides full power
redundancy.
Voltage regulation
Voltage regulators convert the 12V DC supplied from the server power supply into a variety of lower
voltages used by the different system components. HP uses voltage regulators that maintain greater
than 90% efficiency over a broad range of power requirements. The net result is near an 8% gain in
DC power efficiency, which results in almost 10% efficiency gain in AC input power. These efficiency
gains come with no loss in performance and require no configuration by the user.
Improved thermal sensors and fan control
The ProLiant AMD-based 300-series G7 servers include many more thermal sensors —referred to as a
“sea of sensors” — these sensors are located on DIMMs, hard drives, and elsewhere throughout the
server. The actual number of sensors varies by server platform.
The previous generation of ProLiant servers marked a shift away from processors as the primary
producers of heat in the server. As memory modules become denser, they generate more heat. To
combat this, DDR3 DIMMs, as used in the ProLiant G7 servers, incorporate the first reliable on-DIMM
thermal sensors
Because hard drive thermal sensors were not directly associated with fans, the fans would often
operate at high speeds to prevent hard drives from overheating. ProLiant 300-series G7 servers
incorporate hard drive temperature sensors into the body of data used to determine fan speed. This
requires collaboration among various pieces of firmware, including the iLO firmware, system
firmware, and RAID storage controller firmware. The 300-series G7 servers have “zoned” fans that
increase cooling and energy efficiencies in the server by adjusting cooling to those zones when called
for by the sensors. This provides improved efficiency and better acoustics for the platform. The iLO
management processor in the G7 300 series uses a sophisticated control algorithm to set the speed
for each fan zone in the system based on feedback from the appropriate temperature sensors. This
allows fans to consume the minimum amount of required power.
The fan control algorithm lets ProLiant 300-series G7 servers change fan speed as the situation
dictates. In ProLiant AMD-based servers prior to G7, if one fan failed, all the other fans were set to
high speed to assure the server remained within thermal specifications. ProLiant 300-series G7 servers
12
include enough sensors to construct an accurate view of the thermal landscape within the server,
allowing the sensors and the fan control algorithm to determine if fan speeds need to be increased.
HP Dynamic Power Capping
Because Dynamic Power Capping lets you keep server power consumption below a power threshold
in real time, you can use it as a tool for planning and managing both electrical provisioning and
cooling requirements in the data center. You can electrically provision a PDU or a rack to less than
the full faceplate power rating of all the servers loaded and be assured it will not exceed the set limits.
To implement Dynamic Power Capping, the iLO management processor works in conjunction with a
power microcontroller both to measure and to control power use. When enforcing the Dynamic Power
Cap, the power microcontroller keeps the processor’s performance and power use under the set cap.
This process is illustrated in Figure 4. You can set a Dynamic Power Cap for an individual server from
the iLO Advanced user interface. For multiple rack-mount servers, you can set the Dynamic Power
Caps from the power management module within HP Insight Control Environment.
HP Dynamic Power Capping is operating system independent and functions with all operating systems
and software applications. HP Dynamic Power Capping will continue to function even if the software
fails because HP designed the hardware to be independent of the OS.
Since Dynamic Power Capping can impact server performance if set too aggressively, HP
recommends that Dynamic Power Caps be set at values that match or exceed the highest observed
power consumption over a representative server workload sample.
For a more detailed explanation of HP Dynamic Power Capping, see the technology brief “HP Power
Capping and HP Dynamic Power Capping for ProLiant servers” available at
http://h20000.www2.hp.com/bc/docs/support/SupportManual/c01549455/c01549455.pdf.
13
Figure 4. Rapid response of Dynamic Power Capping avoids circuit breaker trips
Managing processor technologies
HP leverages several of the power management technologies built into AMD Opteron 6100 series
processor. This suite of technologies includes AMD PowerNow™ Technology, enhanced Advanced
Platform Management Link (APML), and AMD Power Cap Manager™.
APML Remote Power Management Interface
The AMD Opteron six-core 6100-series processors incorporate multiple thermal sensors to indicate the
hottest part of the processor. Using a systems management device such as iLO, you can remotely
monitor and control P-state limits using AMD’s APML Remote Power Management Interface. The iLO
processor includes the processor thermal data in the data it sends to the fan controller.
Enhanced APML
Enhanced APML is a new capability offered on HP ProLiant G7 platforms with the AMD 6100 series
processors. This capability includes a Precision Thermal Monitor feature that provides a more precise
digital readout of CPU thermals, monitors power/cooling, and alerts the Baseboard Management
Controller (BMC). HP ProLiant thermal policy relies on enhanced APML for accurate processor thermal
information.
AMD CoolCore™
This AMD technology reduces processor energy consumption by turning off unused parts of the
processor. This in turn helps reduce power and cooling costs for the IT infrastructure. AMD CoolCore
is enabled by default and is not a user option.
14
Cool Speed ™
Cool Speed technology protects processor integrity by reducing power-states when the processor
reaches an established temperature limit. This technology allows a server to continue operating if the
processor thermal environment exceeds safe operational limits.
It is important for managers to understand that Cool Speed is an automated processor feature. Within
a rack environment, some server platforms may initiate this process while other servers do not. Given
the same workload, Cool Speed activation will depend on the installed processor model and local
environmental conditions. AMD Cool Speed is enabled by default and is not a user option.
C1E ™
In G34 based-systems, the Northbridge chipset detects when all of the processor cores are idle and
communicates that command to the Southbridge chipset. After the command is sent, the Northbridge
and HT links are powered down and the cores go into a deeper sleep state. Depending on system
configuration, this feature can equate to significant data center power savings when the Northbridge
and HT links are powered down and cores are at idle. The OS manages C1E and it is enabled or
disabled through the RBSU.
PowerNow! Technology
AMD PowerNow! Technology with Independent Dynamic Core technology and Dual Dynamic Power
Management™ is technology that allows the processors to run dynamically at different frequencies
and voltages depending upon CPU computing demand. As a result, PowerNow! can lower server
power consumption without compromising performance.
AMD PowerNow! can be enabled on ProLiant AMD-based 300-series G7 servers through the BIOScontrolled Dynamic Mode of Power Regulator for ProLiant, which does not require an OS driver.
Power Cap Manager
Power Cap Manager allows administrators to set a fixed limit on a server's processor power
consumption. In a multi-core environment, administrators control the P-state of individual cores. The
caveat is that a single voltage is supplied to all cores in the processor even if different P-states are
requested by the user. If an administrator requests different P-states for different cores in the processor,
the actual P-state will equal that of the highest voltage required among the selected P-states.
Administrators can control this through HP Dynamic Power Capping, accessed from iLO advanced
menus or the HP ICE management suite.
AMD Core Select
AMD Core Select lets administrators use the RBSU to select the number of software-visible cores per
CPU (minimum of one, up to full number supported). Operating systems and applications can
recognize the reduced core count so that fewer software licenses may be required. By reducing core
count recognition, Core Select has the potential to reduce software licensing costs. Also, applications
can benefit from increased memory bandwidth and cache per core. This feature is offered on HP
ProLiant G7 platforms with the AMD 6100 series processors. The feature is also supported on ProLiant
G6 servers that are configured with the AMD 2400 or 8400 series processors and the latest System
ROM supporting this feature.
For more information on these processor management technologies, please consult the HP ROMBased Setup Utility User Guide for more details on these options.
http://h20000.www2.hp.com/bc/docs/support/SupportManual/c00191707/c00191707.pdf
15
Systems management and monitoring
HP offers management tools to program and control all aspects of the dynamic server environment.
The HP iLO 3 management processor provides remote management with other core-embedded
management functions to simplify setup, health monitoring, power and thermal control, and remote
administration. The HP Insight Control Environment management suite provides a foundation for
deploying, managing, optimizing, and controlling the entire server environment from any location. HP
Insight Dynamics for ProLiant delivers comprehensive functions for optimizing and balancing resources
and workloads in real time.
HP Integrated Lights-Out 3
ProLiant 300-series G7 servers include the next generation of lights-out management, iLO 3. HP
developed iLO 3 to improve performance, streamline the user interface, and enhance support for
management standards. Like iLO 2, the iLO3 firmware can be upgraded with HP iLO Advanced for
extended functionality.
The iLO 3 hardware includes a faster processor along with increased flash and DDR memory for
storing and executing programs. Remote console performance is up to eight times faster than iLO2—
and equal to the performance of KVM and software-based remote management solutions. The remote
console for Microsoft Windows in iLO 3 is built on the .NET framework using a DirectX control; the
console display scales without scroll bars, supports resolutions up to 1600X1200, and supports
display across multiple monitors. Users have direct access to the remote console without having to
navigate through the iLO 3 web interface. An integrated Linux version provides a single java applet
to support remote console, virtual media and virtual power functionality. The faster iLO 3 processor
and a USB 2.0 host connection also enable three times better virtual media performance compared to
iLO 2.
The web-based graphical user interface has been updated with a look and feel similar to the
Onboard Administrator; it uses Javascript Object Notation (JSON) for improved handling of dynamic
content.
For environments with elevated security requirements, the iLO 3 ASIC includes an AES encryption
engine. AES (Advanced Encryption Standard) is an encryption standard adopted by the U.S.
government (FIPS PUBS 197, http://csrc.nist.gov/publications/fips/fips197/fips-197.pdf). The iLO 3
hardware-based solution provides strong AES encryption with better performance than a softwarebased solution. The latest generation of iLO also provides IPMI over LAN and DCMI management
interfaces for environments requiring simple standards-based monitoring and control.
Not all iLO 2 features are supported in iLO 3. For detailed information on standard and iLO
Advanced licensed features, see the iLO QuickSpecs available at: www.hp.com/go/iLO.
HP Insight Control Environment
HP Insight Control Environment management suite is the infrastructure management suite that HP
recommends for every HP G6 and G7 server. Insight Control Environment provides infrastructure
management functions, including a complete set of lifecycle management tools.
For more information about HP Insight Control Environment management suite, go to
www.hp.com/go/ice
HP Insight Dynamics
HP provides tools like HP Insight Dynamics for ProLiant to assist you with continuously analyzing and
optimizing your data center infrastructure. HP Insight Dynamics delivers four key capabilities for HP
ProLiant 300-series G7 servers:
16
• Continuous capacity and power planning
• Balanced physical and virtual resources
• Cost-effective availability
• Consistent infrastructure provisioning
For more information about HP Insight Dynamics, go to www.hp.com/go/insightdynamics
Security
The Trusted Platform Module™ (TPM) and Microsoft® BitLocker® technology are supported in all
ProLiant 300-series G7 servers by means of the Trusted Platform Module option kit. The Trusted
Platform Module v1.2 supported on ProLiant G7 servers is a microcontroller chip that can create,
securely store, and manage artifacts such as passwords, certificates, and encryption keys that are
used to authenticate the server platform. The TPM 1.2 chip provides a unique Endorsement Key (EK)
and a unique Storage Root Key (SRK). It provides data encryption and uses RSA, SHA-1, RNG
cryptographic functions to provide access protection, OS level protection, and stolen disk protection.
The TPM 1.2 chip can also store platform measurements (hashes) to help ensure that the platform
remains trustworthy. TPM enables Microsoft BitLocker, part of Windows® Server 2008. TPM is an
option on all ProLiant 100-series G6 and G7 servers. For more information about TPM, go to
www.hp.com/go/TPM
Microsoft BitLocker Drive Encryption (BitLocker) is a data protection feature available in Windows
Server 2008. BitLocker uses the enhanced security capabilities of TPM version 1.2 to protect data and
to ensure that a server running Windows Server 2008 has not been compromised while the system
was offline.
OS support
HP performs extensive testing, qualification, and certification of the latest server operating systems on
ProLiant servers to ensure maximum performance and reliability. HP resells and provides full service
and support for Microsoft Windows operating systems, Red Hat Linux® subscriptions, Novell SUSE
Linux subscriptions, Citrix XenServer, and VMware hypervisors. The latest information regarding
support and deployment can be found online at www.hp.com/go/ossupport.
Summary
The HP ProLiant 300-series G7 servers equipped with AMD processors help administrators increase
business performance, lower power costs, and manage their server hardware more easily. To
improve performance, these servers use the AMD Opteron 6100-series processor technologies with
integrated memory controllers and DDR3 memory. The latest Smart Array controllers and firmware
improve RAID performance compared to the previous generation of controllers. Using HP common slot
power supplies is another means for customers to refine and constrain server power based on their
data center requirements. iLO 3, HP Insight Control Environment management suite, and HP Insight
Dynamics all facilitate server management. These servers can be easily deployed with SmartStart,
Insight Control server deployment, and RBSU.
17
For more information
For additional information, refer to the resources listed below.
Resource description
Web address
Dynamic Power Capping TCO and
Best Practices White Paper
www2.hp.com/v2/GetPDF.aspx/4AA2-3107ENW.pdf
HP Insight Control Environment
www.hp.com/go/ice
HP Network Adapters for ProLiant DL
and ML Servers
http://media.hpvitc.veplatform.com/content/HP_Network_Ada
pters_for_ProLiant_DL_Family_data_sheet_1237839147.pdf
HP ProLiant DL385 G7 Server
QuickSpecs
http://h18004.www1.hp.com/products/quickspecs/13594_n
a/13594_na.html
HP ROM-Based Setup Utility User
Guide
http://h20000.www2.hp.com/bc/docs/support/SupportManu
al/c00191707/c00191707.pdf
HP iLO 3 product information
www.hp.com/go/ilo
HP SAS and SATA technology
www.hp.com/go/serial
HP Smart Array Advanced Pack
http://h18004.www1.hp.com/products/servers/proliantstorag
e/arraycontrollers/smartarray-advanced/index.html
HP Smart Array controllers
www.hp.com/products/smartarray
ISS Technology Communications
briefs:
“HP Power Capping and HP
Dynamic Power Capping for ProLiant
servers”
ISS Technology Communications
briefs:
“Memory technology evolution: an
overview of system memory
technologies”
ISS Technology Communications
briefs:
“Serial Attached SCSI storage
technology”
http://h20000.www2.hp.com/bc/docs/support/SupportManu
al/c01549455/c01549455.pdf
http://h20000.www2.hp.com/bc/docs/support/SupportManu
al/c00256987/c00256987.pdf
http://h20000.www2.hp.com/bc/docs/support/SupportManu
al/c01613420/c01613420.pdf
18
Call to action
Send comments about this paper to TechCom@HP.com
Follow us on Twitter: http://www.twitter.com/HPISSTechComm
© Copyright 2010 Hewlett-Packard Development Company, L.P. The information contained herein is subject to
change without notice. The only warranties for HP products and services are set forth in the express warranty
statements accompanying such products and services. Nothing herein should be construed as constituting an
additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein.
AMD is a trademark of Advanced Micro Devices, Inc.
Microsoft, Windows, and BitLocker are U.S. registered trademarks of Microsoft Corporation.
TC100407TB, June 2010
Open as PDF
Similar pages