Xeon 5600 Vol 1 Datasheet
Xeon 5600 Vol 1 Datasheet
Datasheet, Volume 1
June 2011
1 Introduction ............................................................................................................ 11
1.1 Processor Features ............................................................................................ 12
1.2 Platform Features .............................................................................................. 12
1.3 Terminology ..................................................................................................... 13
1.4 References ....................................................................................................... 15
1.5 Statement of Volatility ....................................................................................... 15
2 Electrical Specifications ........................................................................................... 17
2.1 Processor Signaling ........................................................................................... 17
2.1.1 Intel® QuickPath Interconnect ............................................................... 17
2.1.2 DDR3 Signal Groups.............................................................................. 17
2.1.3 Platform Environmental Control Interface (PECI) ....................................... 17
2.1.4 Processor Sideband Signals .................................................................... 18
2.1.5 System Reference Clock ........................................................................ 18
2.1.6 Test Access Port (TAP) Signals................................................................ 19
2.1.7 Power / Other Signals............................................................................ 19
2.1.8 Reserved or Unused Signals ................................................................... 27
2.2 Signal Group Summary ...................................................................................... 27
2.3 Mixing Processors.............................................................................................. 29
2.4 Flexible Motherboard Guidelines (FMB) ................................................................. 30
2.5 Absolute Maximum and Minimum Ratings ............................................................. 30
2.6 Processor DC Specifications ................................................................................ 31
2.6.1 VCC Overshoot Specifications ................................................................. 35
2.6.2 Die Voltage Validation ........................................................................... 36
2.7 Intel QuickPath Interconnect Specifications ........................................................... 44
2.8 AC Specifications............................................................................................... 47
2.9 Processor AC Timing Waveforms ......................................................................... 54
3 Signal Quality Specifications.................................................................................... 65
3.1 Overshoot/Undershoot Tolerance......................................................................... 65
4 Package Mechanical Specifications .......................................................................... 67
4.1 Package Mechanical Specifications ....................................................................... 67
4.1.1 Package Mechanical Drawing .................................................................. 67
4.1.2 Processor Component Keep-Out Zones .................................................... 70
4.1.3 Package Loading Specifications ............................................................... 70
4.1.4 Package Handling Guidelines .................................................................. 70
4.1.5 Package Insertion Specifications ............................................................. 70
4.1.6 Processor Mass Specification .................................................................. 71
4.1.7 Processor Materials ............................................................................... 71
4.1.8 Processor Markings ............................................................................... 71
5 Land Listing............................................................................................................. 73
5.1 Listing by Land Name ........................................................................................ 73
5.2 Listing by Land Number ..................................................................................... 90
6 Signal Definitions .................................................................................................. 109
6.1 Signal Definitions ............................................................................................ 109
7 Thermal Specifications .......................................................................................... 113
7.1 Package Thermal Specifications ......................................................................... 113
7.1.1 Thermal Specifications......................................................................... 113
7.1.2 Thermal Metrology .............................................................................. 128
7.2 Processor Thermal Features .............................................................................. 128
7.2.1 Processor Temperature........................................................................ 128
Revision
Description Date
Number
1 Introduction
The Intel Xeon processor 5600 series features a range of Thermal Design Power (TDP)
envelopes from 40W TDP up to 130W TDP, and is segmented into multiple platforms:
• 2-Socket Frequency Optimized Server/Workstation Platforms support a 130 W
Thermal Design Power (TDP) SKU and up to 6 core support. These platforms
provide optimal overall performance and reliability, in addition to high-end graphics
support.
• 2-Socket Advanced Server/Workstation Platforms support a 95 W Thermal Design
Power (TDP) SKU. These platforms provide optimal overall performance featuring
up to 6 core support.
• 2-Socket Standard Server/Workstation Platforms support 80 W TDP processor
SKUs supporting up to 6 cores. These platforms provide optimal performance per
watt for rack-optimized platforms.
• Low Power Platforms implement 60 W TDP (up to 6 cores) and 40 W TDP (up to 4
cores) processor SKU’s. These processors are intended for dual-processor server
blades and embedded servers.
• 1-Socket Workstation Platforms support Intel® Xeon® Processor W3680. These
platforms enable a wide range of options for either the performance, power, or cost
sensitive customer.
• Platforms supporting Higher Case Temperature Low-Voltage Processors with 60 W
TDP (up to 6 cores) and 40 W TDP (up to 4 cores). The higher case temperatures
are intended to meet the short-term thermal profile requirements of NEBS Level 3.
These 2-socket processors are ideal for thermally-constrained form factors in
embedded servers, communications and storage markets. Specifications denoted
as LV-60W apply to the Intel® Xeon® Processor L5638. Specifications denoted as
LV-40W apply to the Intel® Xeon® Processor L5618.
Note: All references to “chipset” in this document pertain to the Intel® 5520 chipset and the
Intel® 5500 chipset.
Intel is committed to delivering processors for both server and workstation platforms
that maximize performance while meeting all Intel Quality and Reliability goals. The
product’s reliability assessment is based on a datasheet compliant system and
reference use condition. Intel utilizes a broad set of use condition assumptions (that is,
percentage of time in active vs. inactive operation, non-operating conditions, and the
number of power cycles per year) to ensure proper operation over the life of the
product. The reference use condition differs between workstation and server processor
SKU’s. Implementing processors outside of reference use conditions may affect
reliability performance.
Table 1-1. Intel® Xeon® Processor 5600 Series Feature Set Overview
Feature Intel® Xeon® Processor 5600 Series
Data Transfer Rate (GT/s) Two full-width Intel® QuickPath Interconnect links;
Up to 6.40 GT/s in each direction
The Intel Xeon processor 5600 series support all the existing Streaming SIMD
Extensions 2 (SSE2), Streaming SIMD Extensions 3 (SSE3) and Streaming SIMD
Extensions 4 (SSE4) instructions. Additionally, Intel Xeon processor 5600 series
support Intel® AES New Instructions (Intel® AES-NI).
The Intel Xeon processor 5600 series support Direct Cache Access (DCA). DCA enables
supported I/O adapter to pre-fetch data from memory to the processor cache, thereby
avoiding cache misses and improving application response times.
These processors support a maximum physical address size of 40 bits. Also supported
is IA-32e paging which adds support for 1 GB (230) page size in addition to 2 MB and
4 kB page size support for linear to physical address translation.
New memory subsystem capabilities include Low Voltage DDR3 (DDR3L) DIMM support
for power optimization. The Intel Xeon processor 5600 series also add features to
provide improved manageability of memory channels. The DDR_THERM2# signal has
been added to support high-temperature DIMMs and their 2X refresh requirements.
Intel Xeon processor 5600 series are based on a low-power microarchitecture that
supports operation within various C-states. Additionally, six execution cores and power
management coordination logic are optimized to manage C-state support at both the
execution core and package levels. An Intel Turbo Boost Technology optimization
feature is supported on these processors for improved energy efficiency.
Intel® Trusted Execution Technology (Intel® TXT) is also supported and represents a
set of enhanced hardware components designed to help protect sensitive information
from software-based attacks. Features include capabilities in the microprocessor,
chipset, I/O subsystems, and other platform components. When coupled with suitably
enabled operating systems and applications, Intel TXT helps protect the confidentiality
and integrity of data in the face of increasingly hostile security environment.
1.3 Terminology
A ‘#’ symbol after a signal name refers to an active low signal, indicating a signal is in
the active state when driven to a low voltage level. For example, when RESET# is low,
a reset has been requested.
UI n =t n -t n-1
1.4 References
Platform designers are strongly encouraged to maintain familiarity with the most up-to-
date revisions of processor and platform collateral.
Notes:
1. Document is available publicly at http://www.intel.com.
2 Electrical Specifications
On-die termination (ODT) is included on the processor silicon and terminated to VSS.
Intel chipsets also provide ODT, thus eliminating the need to terminate on the system
board. Figure 2-1 illustrates the active ODT.
TX Signal RX
Signal
The PECI interface operates at a nominal voltage set by VTTD. The set of DC electrical
specifications shown in Table 2-13 is used with devices normally operating from a VTTD
interface supply.
VTTD
Minimum VP
Minimum Valid Input
Hysteresis Signal Range
Maximum VN
PECI Ground
The processor core frequency is configured during reset by using values stored within
the device during manufacturing. The stored value sets the lowest core multiplier at
which the particular processor can operate. If higher speeds are desired, the
appropriate ratio can be configured via the IA32_PERF_CTL MSR (MSR 199h); Bits
[15:0].
Clock multiplying within the processor is provided by the internal phase locked loop
(PLL), which requires a constant frequency BCLK_DP, BCLK_DN input, with exceptions
for spread spectrum clocking. DC specifications for the BCLK_DP, BCLK_DN inputs are
provided in Table 2-14 and AC specifications in Table 2-22. These specifications must
be met while also meeting the associated signal quality specifications outlined in
Section 3.
Processor TAP signal DC specifications can be found in Table 2-18. AC specifications are
located in Table 2-27.
Note: While TDI, TMS and TRST# do not include On-Die Termination (ODT), these signals are
weakly pulled-up via a 1-5 kΩ resistor to VTT.
Note: While TCK does not include ODT, this signal is weakly pulled-down via a
1-5 kΩ resistor to VSS.
Table 2-1 outlines the required voltage supplies necessary to support Intel Xeon
processor 5600 series.
VCCPLL 1.80 V Each processor includes dedicated VCCPLL and PLL circuits.
Note:
1. Refer to Table 2-8 for voltage and current specifications.
• 210 each VCC (271 ea. VSS) lands must be supplied with the voltage determined by
the VID[7:0] signals. Table 2-2 defines the voltage level associated with each core
VID pattern. Table 2-9 and Figure 2-3 represent VCC static and transient limits.
• 3 each VCCPLL lands, connected to a 1.8 V supply, power the Phase Lock Loop (PLL)
clock generation circuitry. An on-die PLL filter solution is implemented within the
processor.
• 45 each VDDQ (17 ea. VSS) lands, connected to a 1.50 / 1.35 V supply, provide
power to the processor DDR3 interface. This supply also powers the DDR3 memory
subsystem.
• 7 each VTTA (5 ea. VSS) and 26 ea. VTTD (17 ea. VSS) lands must be supplied with
the voltage determined by the VTT_VID[4:2] signals. Coupled with a 20 mV offset,
this corresponds to a VTT_VID pattern of ‘010xxx10’. Table 2-4 specifies the
voltage levels associated with each VTT_VID pattern. Table 2-11 and Figure 2-10
represent VTT static and transient limits.
All VCC, VCCPLL, VDDQ, VTTA, and VTTD lands must be connected to their respective
processor power planes, while all VSS lands must be connected to the system ground
plane.
Individual processor VID values may be calibrated during manufacturing such that two
devices at the same core frequency may have different default VID settings.
The processor uses eight voltage identification signals, VID[7:0], to support automatic
selection of power supply voltages. Table 2-2 specifies the voltage level corresponding
to the state of VID[7:0]. A ‘1’ in this table refers to a high voltage level and a ‘0’ refers
to a low voltage level. If the processor socket is empty (SKTOCC# pulled high), or the
voltage regulation circuit cannot supply the voltage that is requested, the voltage
regulator must disable itself.
The processor provides the ability to operate while transitioning to an adjacent VID and
its associated processor core voltage (VCC). This is represented by a DC shift in the
loadline. It should be noted that a low-to-high or high-to-low voltage state change may
result in as many VID transitions as necessary to reach the target core voltage.
Transitions above the maximum specified VID are not permitted. Table 2-8 includes VID
step sizes and DC shift ranges. Minimum and maximum voltages must be maintained
as shown in Table 2-9.
The VRM or EVRD utilized must be capable of regulating its output to the value defined
by the new VID. DC specifications for dynamic VID transitions are included in
Table 2-18, while AC specifications are included in Table 2-28.
0 0 0 0 0 0 0 0 OFF
0 0 0 0 0 0 0 1 OFF
0 0 0 0 0 0 1 0 1.60000
0 0 0 0 0 0 1 1 1.59375
0 0 0 0 0 1 0 0 1.58750
0 0 0 0 0 1 0 1 1.58125
0 0 0 0 0 1 1 0 1.57500
0 0 0 0 0 1 1 1 1.56875
0 0 0 0 1 0 0 0 1.56250
0 0 0 0 1 0 0 1 1.55625
0 0 0 0 1 0 1 0 1.55000
0 0 0 0 1 0 1 1 1.54375
0 0 0 0 1 1 0 0 1.53750
0 0 0 0 1 1 0 1 1.53125
0 0 0 0 1 1 1 0 1.52500
0 0 0 0 1 1 1 1 1.51875
0 0 0 1 0 0 0 0 1.51250
0 0 0 1 0 0 0 1 1.50625
0 0 0 1 0 0 1 0 1.50000
0 0 0 1 0 0 1 1 1.49375
0 0 0 1 0 1 0 0 1.48750
0 0 0 1 0 1 0 1 1.48125
0 0 0 1 0 1 1 0 1.47500
0 0 0 1 0 1 1 1 1.46875
0 0 0 1 1 0 0 0 1.46250
0 0 0 1 1 0 0 1 1.45625
0 0 0 1 1 0 1 0 1.45000
0 0 0 1 1 0 1 1 1.44375
0 0 0 1 1 1 0 0 1.43750
0 0 0 1 1 1 0 1 1.43125
0 0 0 1 1 1 1 0 1.42500
0 0 0 1 1 1 1 1 1.41875
0 0 1 0 0 0 0 0 1.41250
0 0 1 0 0 0 0 1 1.40625
0 0 1 0 0 0 1 0 1.40000
0 0 1 0 0 0 1 1 1.39375
0 0 1 0 0 1 0 0 1.38750
0 0 1 0 0 1 0 1 1.38125
0 0 1 0 0 1 1 0 1.37500
0 0 1 0 0 1 1 1 1.36875
0 0 1 0 1 0 0 0 1.36250
0 0 1 0 1 0 0 1 1.35625
0 0 1 0 1 0 1 0 1.35000
0 0 1 0 1 0 1 1 1.34375
0 0 1 0 1 1 0 0 1.33750
0 0 1 0 1 1 0 1 1.33125
0 0 1 0 1 1 1 0 1.32500
0 0 1 0 1 1 1 1 1.31875
0 0 1 1 0 0 0 0 1.31250
0 0 1 1 0 0 0 1 1.30625
0 0 1 1 0 0 1 0 1.30000
0 0 1 1 0 0 1 1 1.29375
0 0 1 1 0 1 0 0 1.28750
0 0 1 1 0 1 0 1 1.28125
0 0 1 1 0 1 1 0 1.27500
0 0 1 1 0 1 1 1 1.26875
0 0 1 1 1 0 0 0 1.26250
0 0 1 1 1 0 0 1 1.25625
0 0 1 1 1 0 1 0 1.25000
0 0 1 1 1 0 1 1 1.24375
0 0 1 1 1 1 0 0 1.23750
0 0 1 1 1 1 0 1 1.23125
0 0 1 1 1 1 1 0 1.22500
0 0 1 1 1 1 1 1 1.21875
0 1 0 0 0 0 0 0 1.21250
0 1 0 0 0 0 0 1 1.20625
0 1 0 0 0 0 1 0 1.20000
0 1 0 0 0 0 1 1 1.19375
0 1 0 0 0 1 0 0 1.18750
0 1 0 0 0 1 0 1 1.18125
0 1 0 0 0 1 1 0 1.17500
0 1 0 0 0 1 1 1 1.16875
0 1 0 0 1 0 0 0 1.16250
0 1 0 0 1 0 0 1 1.15625
0 1 0 0 1 0 1 0 1.15000
0 1 0 0 1 0 1 1 1.14375
0 1 0 0 1 1 0 0 1.13750
0 1 0 0 1 1 0 1 1.13125
0 1 0 0 1 1 1 0 1.12500
0 1 0 0 1 1 1 1 1.11875
0 1 0 1 0 0 0 0 1.11250
0 1 0 1 0 0 0 1 1.10625
0 1 0 1 0 0 1 0 1.10000
0 1 0 1 0 0 1 1 1.09375
0 1 0 1 0 1 0 0 1.08750
0 1 0 1 0 1 0 1 1.08125
0 1 0 1 0 1 1 0 1.07500
0 1 0 1 0 1 1 1 1.06875
0 1 0 1 1 0 0 0 1.06250
0 1 0 1 1 0 0 1 1.05625
0 1 0 1 1 0 1 0 1.05000
0 1 0 1 1 0 1 1 1.04375
0 1 0 1 1 1 0 0 1.03750
0 1 0 1 1 1 0 1 1.03125
0 1 0 1 1 1 1 0 1.02500
0 1 0 1 1 1 1 1 1.01875
0 1 1 0 0 0 0 0 1.01250
0 1 1 0 0 0 0 1 1.00625
0 1 1 0 0 0 1 0 1.00000
0 1 1 0 0 0 1 1 0.99375
0 1 1 0 0 1 0 0 0.98750
0 1 1 0 0 1 0 1 0.98125
0 1 1 0 0 1 1 0 0.97500
0 1 1 0 0 1 1 1 0.96875
0 1 1 0 1 0 0 0 0.96250
0 1 1 0 1 0 0 1 0.95625
0 1 1 0 1 0 1 0 0.95000
0 1 1 0 1 0 1 1 0.94375
0 1 1 0 1 1 0 0 0.93750
0 1 1 0 1 1 0 1 0.93125
0 1 1 0 1 1 1 0 0.92500
0 1 1 0 1 1 1 1 0.91875
0 1 1 1 0 0 0 0 0.91250
0 1 1 1 0 0 0 1 0.90625
0 1 1 1 0 0 1 0 0.90000
0 1 1 1 0 0 1 1 0.89375
0 1 1 1 0 1 0 0 0.88750
0 1 1 1 0 1 0 1 0.88125
0 1 1 1 0 1 1 0 0.87500
0 1 1 1 0 1 1 1 0.86875
0 1 1 1 1 0 0 0 0.86250
0 1 1 1 1 0 0 1 0.85625
0 1 1 1 1 0 1 0 0.85000
0 1 1 1 1 0 1 1 0.84375
0 1 1 1 1 1 0 0 0.83750
0 1 1 1 1 1 0 1 0.83125
0 1 1 1 1 1 1 0 0.82500
0 1 1 1 1 1 1 1 0.81875
1 0 0 0 0 0 0 0 0.81250
1 0 0 0 0 0 0 1 0.80625
1 0 0 0 0 0 1 0 0.80000
1 0 0 0 0 0 1 1 0.79375
1 0 0 0 0 1 0 0 0.78750
1 0 0 0 0 1 0 1 0.78125
1 0 0 0 0 1 1 0 0.77500
1 0 0 0 0 1 1 1 0.76875
1 0 0 0 1 0 0 0 0.76250
1 0 0 0 1 0 0 1 0.75625
1 0 0 0 1 0 1 0 0.75000
1 0 0 0 1 0 1 1 0.74375
1 0 0 0 1 1 0 0 0.73750
1 0 0 0 1 1 0 1 0.73125
1 0 0 0 1 1 1 0 0.72500
1 0 0 0 1 1 1 1 0.71875
1 0 0 1 0 0 0 0 0.71250
1 0 0 1 0 0 0 1 0.70625
1 0 0 1 0 0 1 0 0.70000
1 0 0 1 0 0 1 1 0.69375
1 0 0 1 0 1 0 0 0.68750
1 0 0 1 0 1 0 1 0.68125
1 0 0 1 0 1 1 0 0.67500
1 0 0 1 0 1 1 1 0.66875
1 0 0 1 1 0 0 0 0.66250
1 0 0 1 1 0 0 1 0.65625
1 0 0 1 1 0 1 0 0.65000
1 0 0 1 1 0 1 1 0.64375
1 0 0 1 1 1 0 0 0.63750
1 0 0 1 1 1 0 1 0.63125
1 0 0 1 1 1 1 0 0.62500
1 0 0 1 1 1 1 1 0.61875
1 0 1 0 0 0 0 0 0.61250
1 0 1 0 0 0 0 1 0.60625
1 0 1 0 0 0 1 0 0.60000
1 0 1 0 0 0 1 1 0.59375
1 0 1 0 0 1 0 0 0.58750
1 0 1 0 0 1 0 1 0.58125
1 0 1 0 0 1 1 0 0.57500
1 0 1 0 0 1 1 1 0.56875
1 0 1 0 1 0 0 0 0.56250
1 0 1 0 1 0 0 1 0.55625
1 0 1 0 1 0 1 0 0.55000
1 0 1 0 1 0 1 1 0.54375
1 0 1 0 1 1 0 0 0.53750
1 0 1 0 1 1 0 1 0.53125
1 0 1 0 1 1 1 0 0.52500
1 0 1 0 1 1 1 1 0.51875
1 0 1 1 0 0 0 0 0.51250
1 0 1 1 0 0 0 1 0.50625
1 0 1 1 0 0 1 0 0.50000
1 1 1 1 1 1 1 0 OFF
1 1 1 1 1 1 1 1 OFF
Notes:
1. When the “11111111” VID pattern is observed, or when the SKTOCC# pin is pulled high, the voltage
regulator output should be disabled.
2. The VID range includes VID transitions that may be initiated by thermal events, Extended HALT state
transitions (see Section 8.2), higher C-States (see Section 8.2) or Enhanced Intel SpeedStep® Technology
transitions (see Section 8.5). The Extended HALT state must be enabled for the processor to
remain within its specifications
3. Once the VRM/EVRD is operating after power-up, if either the Output Enable signal is de-asserted or a
specific VID off code is received, the VRM/EVRD must turn off its output (the output should go to high
impedance) within 500 ms and latch off until power is cycled.
Note:
1. This setting is defined for future use; no Intel Xeon processor 5600 series SKU is defined with ICC_MAX=40
A.
2. In general, set PWM IMON slope to 900 mV = IMAX, where IMAX = ICCMAX. For the 130 W SKU, set IMON
slope to 900 mV= 180 A. All other SKUs must match the values shown above. Please consult the PWM
datasheet for the IMON slope setting.
Some POC signals include specific timing requirements. Please refer to Section 8.1 for
further details.
Individual processor VTT_VID values may be calibrated during manufacturing such that
two devices at the same core frequency may have different default VTT_VID settings.
The processor utilizes three voltage identification signals to support automatic selection
of power supply voltages. These correspond to VTT_VID[4:2]. The VTT voltage level
delivered to the processor lands must also encompass a 20 mV offset (See Table 2-4;
VTT_TYP) above the voltage level corresponding to the state of the VTT_VID[7:0] signals
(See Table 2-4; VR 11.0 Voltage). Table 2-11 and Figure 2-10 provide the resulting
static and transient tolerances. Please note that the maximum and minimum electrical
loadlines are defined by a 31.5 mV tolerance band above and below VTT_TYP values.
0 1 0 0 0 0 1 0 1.200 V 1.220 V
0 1 0 0 0 1 1 0 1.175 V 1.195 V
0 1 0 0 1 0 1 0 1.150 V 1.170 V
0 1 0 0 1 1 1 0 1.125 V 1.145 V
0 1 0 1 0 0 1 0 1.100 V 1.120 V
0 1 0 1 0 1 1 0 1.075 V 1.095 V
0 1 0 1 1 0 1 0 1.050 V 1.070 V
0 1 0 1 1 1 1 0 1.025 V 1.045 V
TAP signals do not include on-die termination, however they may include resistors on
package (refer to Section 2.1.6 for details). Inputs and utilized outputs must be
terminated on the baseboard. Unused outputs may be terminated on the baseboard or
left unconnected. Note that leaving unused outputs unterminated may interfere with
some TAP functions, complicate debug probing, and prevent boundary scan testing.
PWRGOOD Signal
Reset Signal
Power/Other Signals
Notes:
1. Refer to Section 5 for land assignments and Section 6 for signal definitions.
2. DDR{0/1/2} refers to DDR3 channel 0, DDR3 channel 1, and DDR3 Channel 2
Signals that include on-die termination (ODT) are listed in Table 2-6.
Notes:
1. Unless otherwise specified, signals have ODT in the package with a 50 Ω pull-down to VSS.
2. Unless otherwise specified, all DDR3 signals are terminated to VDDQ/2.
3. DDR{0/1/2}_PAR_ERR#[2:0] are terminated to VDDQ.
4. TCK does not include ODT, this signal is weakly pulled-down via a 1-5 kΩ resistor to VSS.
5. TDI, TMS, TRST# do not include ODT, these signals are weakly pulled-up via 1-5kΩ resistor to VTT.
6. BPM[7:0]# and PREQ# signals have ODT in package with 35 Ω pull-ups to VTT.
7. PECI_ID# has ODT in package with a 1-5 kΩ pull-up to VTT.
8. TAPPWRGOOD has ODT in package with a 1-2.5 kΩ pull-up to VTT.
9. VCCPWRGOOD, VDDPWRGOOD, and VTTPWRGOOD have ODT in package with a 5-20 kΩ pull-down to VSS.
Note: Processors must operate with the same Intel QuickPath Interconnect, DDR3 memory
and core frequency.
While Intel does nothing to prevent processors from operating together, some
combinations may not be supported due to limited validation, which may result in
uncharacterized errata. Coupling this fact with the large number of Intel Xeon
processor 5600 series processor attributes, the following population rules and stepping
matrix have been developed to clearly define supported configurations.
• Processors must be of the same power-optimization segment. This insures
processors include the same maximum Intel QuickPath Interconnect and DDR3
operating speeds and cache sizes.
• Processors must operate at the same core frequency. Note, processors within the
same power-optimization segment supporting different maximum core frequencies
can be operated within a system. However, both must operate at the highest
frequency rating commonly supported. Mixing components operating at different
internal clock frequencies is not supported and will not be validated by Intel.
• Processors must share symmetry across physical packages with respect to the
number of logical processors per package, number of cores per package (but not
necessarily the same subset of cores within the packages), number of Intel
QuickPath Interconnect interfaces and cache topology.
• Mixing steppings is only supported with processors that have identical Extended
Family, Extended Model, Processor Type, Family Code and Model Number as
indicated by the Function 1 of the CPUID instruction. Mixing processors of different
steppings, but the same mode (as per CPUID instruction) is supported. Details
regarding the CPUID instruction are provide in the Intel® 64 and IA-32
Architectures Software Developer’s Manual, Volume 2A.
• After AND’ing the feature flag and extended feature flags from the installed
processors, any processor whose set of feature flags exactly matches the AND’ed
feature flags can be selected by the BIOS as the BSP. If no processor exactly
matches the AND’ed feature flag values, then the processor with the numerically
lower CPUID should be selected as the BSP.
• Intel requires that the proper microcode update be loaded on each processor
operating within the system. Any processor that does not have the proper
microcode update loaded is considered by Intel to be operating out-of-specification.
• Customers are fully responsible for the validation of their system configurations
Note: Processors within a system must operate at the same frequency per bits [15:8] of the
FLEX_RATIO MSR (Address: 194h); however this does not apply to frequency
transitions initiated due to thermal events, Extended HALT, Enhanced Intel SpeedStep
Technology transitions signal (See Section 8).
Although the processor contains protective circuitry to resist damage from static
electric discharge, precautions should always be taken to avoid high static voltages or
electric fields.
Notes:
1. For functional operation, all processor electrical, signal quality, mechanical and thermal specifications must
be satisfied.
2. Overshoot and undershoot voltage guidelines for input, output, and I/O signals are outlined in Section 3.
Excessive overshoot or undershoot on any signal will likely result in permanent damage to the processor.
3. VTTA and VTTD should be derived from the same voltage regulator (VR).
4. 5% tolerance
5. Storage temperature is applicable to storage conditions only. In this scenario, the processor must not
receive a clock, and no lands can be connected to a voltage bias. Storage within these limits will not affect
the long-term reliability of the device. For functional operation, please refer to the processor case
temperature specification.
6. This rating applies to the processor and does not include any tray or packaging.
7. Failure to adhere to this specification can affect the long-term reliability of the processor
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processors.
2. Individual processor VID and/or VTT_VID values may be calibrated during manufacturing such that two
devices at the same speed may have different settings.
3. These voltages are targets only. A variable voltage source should exist on systems in the event that a
different voltage is required.
4. The VCC voltage specification requirements are measured across vias on the platform for the VCCSENSE and
VSSSENSE pins close to the socket with a 100 MHz bandwidth oscilloscope, 1.5 pF maximum probe
capacitance, and 1 MΩ minimum impedance. The maximum length of ground wire on the probe should be
less than 5 mm. Ensure external noise from the system is not coupled in the scope probe.
5. The VTT voltage specification requirements are measured across vias on the platform for the VTTD_SENSE
and VSS_SENSE_VTTD lands close to the socket with a 100 MHz bandwidth oscilloscope, 1.5 pF maximum
probe capacitance, and 1 MΩ minimum impedance. The maximum length of ground wire on the probe should
be less than 5 mm. Ensure external noise from the system is not coupled in the scope probe.
6. Refer to Table 2-9 and corresponding Figure 2-3. The processor should not be subjected to any static VCC
level that exceeds the VCC_MAX associated with any particular current. Failure to adhere to this specification
can shorten processor lifetime.
7. Minimum VCC and maximum ICC are specified at the maximum processor case temperature (TCASE) shown in
Table 7-1. ICC_MAX is specified at the relative VCC_MAX point on the VCC load line. The processor is capable of
drawing ICC_MAX for up to 10 ms. Refer to Figure 2-5 through Figure 2-8 for further details on the average
processor current draw over various time durations.
8. Refer to Table 2-11 and corresponding Figure 2-10. The processor should not be subjected to any static VTT
level that exceeds the VTT_MAX associated with any particular current. Failure to adhere to this specification
can shorten processor lifetime.
9. This specification represents the VCC reduction due to each VID transition. See Section 2.1.7.3. AC timing
requirements for VID transitions are included in Figure 2-29.
10.Baseboard bandwidth is limited to 20 MHz.
11.FMB is the flexible motherboard guidelines. See Section 2.4 for FMB details.
12.ICC_TDC (Thermal Design Current) is the sustained (DC equivalent) current that the processor is capable of
drawing indefinitely and should be used for the voltage regulator temperature assessment. The voltage
regulator is responsible for monitoring its temperature and asserting the necessary signal to inform the
processor of a thermal excursion.
Notes:
1. The VCC_MIN and VCC_MAX loadlines represent static and transient limits. Please see Section 2.6.1 for VCC
overshoot specifications.
2. This table is intended to aid in reading discrete points on Figure 2-3.
3. The loadlines specify voltage limits at the die measured at the VCC_SENSE and VSS_SENSE lands. Voltage
regulation feedback for voltage regulator circuits must also be taken from processor VCC_SENSE and
VSS_SENSE lands.
4. Processor core current (ICC) ranges are valid up to ICC_MAX of the processor SKU as defined in Table 2-8.
Icc [A]
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
VID - 0.000
VID - 0.020
VID - 0.040
VID - 0.060
VID - 0.080
Vcc [V]
VID - 0.100
VID - 0.120
VID - 0.140
VID - 0.160
VID - 0.180
Notes:
1. The VCC_MIN and VCC_MAX loadlines represent static and transient limits. Please see Section 2.6.1 for VCC
overshoot specifications.
2. Refer to Table 2-9 for VCC Static and Transient Tolerance.
3. The loadlines specify voltage limits at the die measured at the VCC_SENSE and VSS_SENSE lands. Voltage
regulation feedback for voltage regulator circuits must also be taken from processor VCC_SENSE and
VSS_SENSE lands.
4. Processor core current (ICC) ranges are valid up to ICC_MAX of the processor SKU as defined in Table 2-8.
VID - 0.000
TOS
0 5 10 15 20 25
Time [us]
Notes:
1. VOS is the measured overshoot voltage.
2. TOS is the measured time duration above VID.
155
150
145
140
Sustained Current (A)
135
130
125
120
115
110
105
0.01 0.1 1 10 100 1000
Time Duration, (s)
Notes:
1. Processor or voltage regulator thermal protection circuitry should not trip for load currents greater than
ICC_TDC.
2. Not 100% tested. Specified by design characterization.
125.0
120.0
Sustained Current (A)
115.0
110.0
105.0
100.0
95.0
0.01 0.1 1 10 100 1000
Time Duration, (s)
Notes:
1. Processor or voltage regulator thermal protection circuitry should not trip for load currents greater than
ICC_TDC.
2. Not 100% tested. Specified by design characterization.
105
100
90
85
80
75
70
65
0.01 0.1 1 10 100 1000
Time Duration, (s)
Notes:
1. Processor or voltage regulator thermal protection circuitry should not trip for load currents greater than
ICC_TDC.
2. Not 100% tested. Specified by design characterization.
Figure 2-8. Load Current Versus Time (Low Power & LV-60W)1,2
85
80
75
Sustained Current (A)
70
65
60
55
50
45
0.01 0.1 1 10 100 1000
Time Duration, (s)
Notes:
1. Processor or voltage regulator thermal protection circuitry should not trip for load currents greater than
ICC_TDC.
2. Not 100% tested. Specified by design characterization.
Figure 2-9. Load Current Versus Time (Low Power & LV-40W)1,2
55
50
Sustained Current (A)
45
40
35
0.01 0.1 1 10 100 1000
Notes:
1. Processor or voltage regulator thermal protection circuitry should not trip for load currents greater than
ICC_TDC.
2. Not 100% tested. Specified by design characterization.
Note:
1. ITT listed in this table is the sum of ITTA and ITTD.
2. This table is intended to aid in reading discrete points on Figure 2-10.
3. The VTT_MIN and VTT_MAX loadlines represent static and transient limits. Each is characterized by a ±31.5 mV
offset from VTT_TYP.
4. The loadlines specify voltage limits at the die measured at the VTTD_SENSE and VSS_SENSE_VTTD lands.
Voltage regulation feedback for regulator circuits must also be taken from VTTD_SENSE and
VSS_SENSE_VTTD lands.
ITT [A]
0 5 10 15 20 25
0.0500
0.0375
0.0250
0.0125
0.0000
-0.0125
-0.0250
VTT_VID DevIatIon
-0.0375
-0.0500
-0.0625
-0.0750
-0.0875
-0.1000
-0.1125
-0.1250
-0.1375
-0.1500
-0.1625
-0.1750
-0.1875
-0.2000
-0.2125
Notes:
1. The VTT_MIN and VTT_MAX loadlines represent static and transient limits. Each is characterized by a ±31.5 mV
offset from VTT_TYP.
2. Refer to Table 2-4 for processor VTT_VID information.
3. Refer to Table 2-11 for VTT Static and Transient Tolerance.
Clock Buffer On 21 31 Ω 5
RON
Resistance
Command Buffer On 16 24 Ω 5
RON
Resistance
Control Buffer On 21 31 Ω 5
RON
Resistance
Data Buffer On 21 33 Ω 5
RON
Resistance
RESET Buffer On 5 53 Ω 5
RON
Resistance
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. VIL is the maximum voltage level at a receiving agent that will be interpreted as a logical low value.
3. VIH is the minimum voltage level at a receiving agent that will be interpreted as a logical high value.
4. VIH and VOH may experience excursions above VDDQ. However, input signal drivers must comply with the
signal quality specifications. Refer to Section 3.
5. This is the pull down driver resistance.
6. RVTT_TERM is the termination on the DIMM and not controlled by the processor. Please refer to the applicable
DIMM datasheet.
7. The minimum and maximum values for these signals are programmable by BIOS to one of the pairs.
8. COMP resistance must be provided on the system board with 1% resistors.
Note:
1. VTTD supplies the PECI interface. PECI behavior does not affect VTTD min/max specifications.
2. It is expected that the PECI driver will take into account, the variance in the receiver input thresholds and
consequently, be able to drive its output within safe limits (-0.150 V to 0.275*VTTD for the low level and
0.725*VTTD to VTTD+0.150 for the high level).
3. The leakage specification applies to powered devices on the PECI bus.
4. One node is counted for each client and one node for the system host. Extended trace lengths might appear
as additional nodes.
5. Excessive capacitive loading on the PECI line may slow down the signal rise/fall times and consequently
limit the maximum bit rate at which the interface can operate.
6. Please refer to Figure 2-2 for further information.
Absolute 2-16
Vcross (abs) 0.250 0.550 V 2, 4
Crossing Point 2-19
Range of
ΔVcross Crossing Points
N/A 0.140 V 2-20 6
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. Crossing Voltage is defined as the instantaneous voltage value when the rising edge of BCLK_DN is equal to
the falling edge of BCLK_DP.
3. VHavg is the statistical average of the VH measured by the oscilloscope.
4. The crossing point must meet the absolute and relative crossing point specifications simultaneously.
5. VHavg can be measured directly using “Vtop” on Agilent* and “High” on Tektronix* oscilloscopes.
6. VCROSS is defined as the total variation of all crossing voltages as defined in Note 2.
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. The VTTA referred to in these specifications refers to instantaneous VTTA.
3. For VIN between 0 V and VTTA. Measured when the driver is tri-stated.
4. VIH may experience excursions above VTT. However, input signal drivers must comply with the signal quality
specifications in Section 3.
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. The VTTA referred to in these specifications refers to instantaneous VTTA.
3. For VIN between 0 V and VTTA. Measured when the driver is tri-stated.
4. VIH and VOH may experience excursions above VTT. However, input signal drivers must comply with the
signal quality specifications in Section 3.
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. The VTTA referred to in these specifications refers to instantaneous VTTA.
3. For VIN between 0 V and VTTA. Measured when the driver is tri-stated.
4. VIH may experience excursions above VTT. However, input signal drivers must comply with the signal quality
specifications in Section 3.
5. This specification applies to the VCCPWRGOOD and VTTPWRGOOD signals.
6. This specification applies to the VDDPWRGOOD signal.
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. The VTTA referred to in these specifications refers to instantaneous VTTA.
3. RSYS_TERM is the termination on the system and is not controlled by the processor.
4. Applies to all Processor Sideband signals, unless otherwise mentioned in Table 2-6.
5. This specification only applies to DDR_THERM# and DDR_THERM2#signals.
6. COMP resistance must be provided on the system board with 1% resistors. COMP0 resistors are tied to VSS.
BERLane Bit Error Rate per lane valid for 4.8, 5.86 1.0E-14 Events
and 6.4 GT/s
Notes:
1. Used during initialization. It is the state of “OFF” condition for the transmitter. That is, when the output driver is disconnected
and only the minimum termination is connected. The link detection resistor is assumed not connected when specifying this
parameter.
2. Used during initialization. It is the state of “OFF” condition for the receiver when only the minimum termination is connected.
Table 2-20. Parameter Values for Intel QuickPath Interconnect Channels at 4.8 GT/s
(Sheet 1 of 2)
Symbol Parameter Min Max Unit Notes
VTx-cm-dc-pin Transmitter output DC common mode, defined as average 0.23 0.27 Fraction of
of VD+ and VD- Use setup of Figure 2-11. VTX-diff-pp-pin
Table 2-20. Parameter Values for Intel QuickPath Interconnect Channels at 4.8 GT/s
(Sheet 2 of 2)
Symbol Parameter Min Max Unit Notes
VTx-cm-ac-pin Transmitter output AC common mode, defined as ((VD+ + -0.0375 0.0375 Fraction of
VD-)/2 - VTx-cm-ac-pin). Use setup of Figure 2-11 and VTX-diff-pp-pin
Figure 2-13 for illustration of AC common mode distribution
and spec limits.
TXduty-pin Average of UI-UI jitter, using setup of Figure 2-11. This -0.078 0.078 UI
appears as bimodal peaks in UI-UI jitter distribution
Figure 2-14.
TXjitUI-UI-1E-7pin UI-UI jitter measured at Tx output pins with 1E-7 -0.085 0.085 UI
probability, using setup of Figure 2-11. Refer to Figure 2-14
for illustration of UI-UI jitter distribution and spec limits
TXjitUI-UI-1E-9pin UI-UI jitter measured at Tx output pins with 1E-9 -0.09 0.09 UI
probability, using setup of Figure 2-11. Refer to Figure 2-14
for illustration of UI-UI jitter distribution and spec limits
TXclk-acc-jit-N_UI-1E-7 P-P accumulated jitter out of any Tx data or clock over 0 <= 0 0.15 UI
n <= N UI where N=12, measured with 1E-7 probability.
Refer to Figure 2-14 for illustration
TXclk-acc-jit-N_UI-1E-9 P-P accumulated jitter out of any Tx data or clock over 0 <= 0 0.17 UI
n <= N UI where N=12, measured with 1E-9 probability.
Refer to Figure 2-14 for illustration
TTx-data-clk-skew-pin Delay of any data lane relative to the clock lane, as -0.4 0.4 UI
measured at Tx output
VRx-cm-dc-pin DC common mode ranges at the Rx input for any data or 145 350 mV
clock channel, defined as average of VD+ and VD-.
VRx-cm-ac-pin AC common mode ranges at the Rx input for any data or -50 50 mV
clock channel, defined as((VD+ + VD-)/2 - VRX-cm-dc-pin).
Refer to Figure 2-13 for illustration.
TRx-margin Measured timing margin during receiver margining with any 0.1 UI
receiver equalizer off or for Tx EQ only based systems
Notes:
1. Refers to routing lengths of 10 - 15 inches (25.4 - 38.1 cm).
2. Refers to routing lengths of 0 - 10 inches (0 - 25.4 cm).
Table 2-21. Parameter Values for Intel QuickPath Interconnect Channel at 5.86 or
6.4 GT/s
Symbol Parameter Min Max Unit Note
TXduty-pin Average of UI-UI jitter, using setup of Figure 2-11. -0.078 0.078 UI
This appears as bimodal peaks in UI-UI jitter
distribution Figure 2-14
TXjitUI-UI-1E-7pin UI-UI jitter measured at Tx output pins with 1E-7 -0.088 0.088 UI
probability, using setup of Figure 2-11. Refer to
Figure 2-14 for illustration of UI-UI jitter distribution
and spec limits
TXjitUI-UI-1E-9pin UI-UI jitter measured at Tx output pins with 1E-9 -0.095 0.095 UI
probability, using setup of Figure 2-11. Refer to
Figure 2-14 for illustration of UI-UI jitter distribution
and spec limits.
TTx-data-clk-skew-pin Delay of any data lane relative to clock lane, as -0.4 0.4 UI
measured at Tx output
VRx-cm-dc-pin DC common mode ranges at the Rx input for any data 145 350 mV
or clock channel, defined as average of VD+ and VD-.
VRx-cm-ac-pin AC common mode ranges at the Rx input for any data –50 50 mV
or clock channel, defined as ((VD+ + VD-)/2 - VRX-cm-
dc-pin). Refer to Figure 2-13 for illustration.
Notes:
1. Refers to routing lengths of 10 - 15 inches (25.4 - 38.1 cm).
2. Refers to routing lengths of 0 - 10 inches (0 - 25.4 cm).
2.8 AC Specifications
AC specifications are defined at the processor pads, unless otherwise noted.
Therefore, proper simulation is the only means to verify proper timing and signal
quality. Care should be taken to read all notes associated with each parameter.
TBCLK-Dutycycle 40 50 60 % 2-17
TBCLK-diff-jit 500 ps 4
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. SSC is Spread Spectrum Clocking. The processor core clock frequency is derived from BCLK. The system
reference clock to processor core clock ratio is determined during initialization as described in
Section 2.1.5.
3. Rise and fall time slopes (V/ns) are measured between +150 mV and -150 mV of the differential output of
reference clock.
4. Phase drift between reference clocks at two connected ports.
Max Min
Latency Timings
tCL – tRCD – tRP CAS Latency – RAS to CAS Delay – 6–6–6 tCK
Pre-charge Command Period
Electrical Characteristics
Clock Timings
TSU + THD DQ Input Setup plus Hold Time to 0.25 * UI ns 2-23 1,2,7
DQS Rising or Falling Edge
Max Min
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies. Timing
specifications only depend on the operating frequency of the memory channel and not the maximum rated
frequency.
2. When the single ended slew rate of the input Data or Strobe signals, within a byte group, are below 1.0
V/ns, the tSU and tHD specifications must be increased by a derating factor. The input single ended slew
rate is measured DC to AC levels; VIL_DC to VIH_AC for rising edges, and VIH_DC to VIL_AC for falling
edges. Use the worse case minimum slew rate measured between Data and Strobe, within a byte group, to
determine the required derating value. No derating is required for single ended slew rates equal to or
greater than 1.0 V/ns.
3. Edge Placement Accuracy (EPA): The silicon contains digital logic that automatically adjusts the timing
relationship between the DDR reference clocks and DDR signals. The BIOS initiates a training procedure
that will place a given signal appropriately within the clock period. The difference in delay between the
signal and clock is accurate to within ±EPA. This EPA includes jitter, skew, within die variation and several
other effects.
4. Data to Strobe read setup and Data from Strobe read hold minimum requirements specified at the
processor pad are determined with the minimum Read DQS/DQS# delay.
5. CWL (CAS Write Latency) is the delay, in clock cycles, between the rising edge of CK where a write
command is referenced and the first rising strobe edge where the first byte of write data is present. The
CWL value is determined by the value of the CL (CAS Latency) setting.
6. The system memory clock outputs are differential (CLK and CLK#), the CLK rising edge is referenced at the
crossing point where CLK is rising and CLK# is falling.
7. The system memory strobe outputs are differential (DQS and DQS#), the DQS rising edge is referenced at
the crossing point where DQS is rising and DQS# is falling, and the DQS falling edge is referenced at the
crossing point where DQS is falling and DQS# is rising.
8. This values specifies the parameter after write leveling, representing the residual error in the controller
after training, and does not include any effects from the DRAM itself.
Table 2-24. DDR3 Electrical Characteristics and AC Specifications at 1066 MT/s (Sheet 1
of 2)
Channel 0
Channel 1
Symbol Parameter Channel 2 Unit Figure Note
Max Min
Latency Timings
tCL – tRCD – tRP CAS Latency – RAS to CAS Delay – 7–7–7 tCK
Pre-charge Command Period 8-8-8
Electrical Characteristics
Table 2-24. DDR3 Electrical Characteristics and AC Specifications at 1066 MT/s (Sheet 2
of 2)
Channel 0
Channel 1
Symbol Parameter Channel 2 Unit Figure Note
Max Min
Clock Timings
TSU + THD DQ Input Setup plus Hold Time to 0.25 * UI ns 2-23 1,2,7
DQS Rising or Falling Edge
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. When the single ended slew rate of the input Data or Strobe signals, within a byte group, are below 1.0
V/ns, the tSU and tHD specifications must be increased by a derating factor. The input single ended slew
rate is measured DC to AC levels; VIL_DC to VIH_AC for rising edges, and VIH_DC to VIL_AC for falling
edges. Use the worse case minimum slew rate measured between Data and Strobe, within a byte group, to
determine the required derating value. No derating is required for single ended slew rates equal to or
greater than 1.0 V/ns.
3. Edge Placement Accuracy (EPA): The silicon contains digital logic that automatically adjusts the timing
relationship between the DDR reference clocks and DDR signals. The BIOS initiates a training procedure
that will place a given signal appropriately within the clock period. The difference in delay between the
signal and clock is accurate to within ±EPA. This EPA includes jitter, skew, within die variation and several
other effects.
4. Data to Strobe read setup and Data from Strobe read hold minimum requirements specified at the
processor pad are determined with the minimum Read DQS/DQS# delay.
5. CWL (CAS Write Latency) is the delay, in clock cycles, between the rising edge of CK where a write
command is referenced and the first rising strobe edge where the first byte of write data is present. The
CWL value is determined by the value of the CL (CAS Latency) setting.
6. The system memory clock outputs are differential (CLK and CLK#), the CLK rising edge is referenced at the
crossing point where CLK is rising and CLK# is falling.
7. The system memory strobe outputs are differential (DQS and DQS#), the DQS rising edge is referenced at
the crossing point where DQS is rising and DQS# is falling, and the DQS falling edge is referenced at the
crossing point where DQS is falling and DQS# is rising.
8. This values specifies the parameter after write leveling, representing the residual error in the controller
afrter training, and does not include any effects from the DRAM itself.
Max Min
tCL – tRCD – tRP CAS Latency – RAS to CAS Delay – 8-8-8 tCK
Pre-charge Command Period 9-9-9
Electrical Characteristics
Max Min
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. When the single ended slew rate of the input Data or Strobe signals, within a byte group, are below 1.0
V/ns, the tSU and tHD specifications must be increased by a derating factor. The input single ended slew
rate is measured DC to AC levels; VIL_DC to VIH_AC for rising edges, and VIH_DC to VIL_AC for falling
edges. Use the worse case minimum slew rate measured between Data and Strobe, within a byte group, to
determine the required derating value. No derating is required for single ended slew rates equal to or
greater than 1.0 V/ns.
3. Edge Placement Accuracy (EPA): The silicon contains digital logic that automatically adjusts the timing
relationship between the DDR reference clocks and DDR signals. The BIOS initiates a training procedure
that will place a given signal appropriately within the clock period. The difference in delay between the
signal and clock is accurate to within ±EPA. This EPA includes jitter, skew, within die variation and several
other effects.
4. Data to Strobe read setup and Data from Strobe read hold minimum requirements specified at the
processor pad are determined with the minimum Read DQS/DQS# delay.
5. CWL (CAS Write Latency) is the delay, in clock cycles, between the rising edge of CK where a write
command is referenced and the first rising strobe edge where the first byte of write data is present. The
CWL value is determined by the value of the CL (CAS Latency) setting.
6. The system memory clock outputs are differential (CLK and CLK#), the CLK rising edge is referenced at the
crossing point where CLK is rising and CLK# is falling.
7. The system memory strobe outputs are differential (DQS and DQS#), the DQS rising edge is referenced at
the crossing point where DQS is rising and DQS# is falling, and the DQS falling edge is referenced at the
crossing point where DQS is falling and DQS# is rising.
8. This values specifies the parameter after write leveling, representing the residual error in the controller
afrter training, and does not include any effects from the DRAM itself.
TCO: Time from BCLK land until signal valid at output 0.5 2.275 ns 11
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. All AC timings for the Asynchronous GTL signals are referenced to the BCLK_P rising edge at Crossing
Voltage (VCROSS). VCCPWRGOOD, VTTPWRGOOD and VDDPWRGOOD are referenced to BCLK_P rising edge
at 0.5 * VTT.
3. These signals may be driven asynchronously.
4. Refer to Section 8 for additional timing requirements for entering and leaving low power states.
5. xxPWRGOOD signal has no edge rate requirement, but edge must be monotonic.
6. VDDPWRGOOD must be asserted no later then VCCPWRGOOD. There is no releationship between
VDDPWRGOOD and VCC ramp.
7. There is no dependency between VDDPWRGOOD and VTTPWRGOOD assertion.
8. VTTPWRGOOD must accurately reflect the state of VTT and must not glitch whenever VTT or VDD is
applied.
9. VTT must read VTTFINAL before VCCPWRGOOD assertion.
10. It may be required to add delay on the board to meet the 1 ms minimum processor requirement.
11. Based on a test load of 50 Ω to VTT.
12. Specified for synchronous signals.
13. Applies to PROCHOT# signal only. Please see Section 2.1.7.3.1 and Section 8.1 for information regarding
Power-On Configuration options.
14. Rise time is measured from 10% to 90% of the final voltage.
Notes:
1. Unless otherwise noted, all specifications in this table apply to all processor frequencies.
2. Not 100% tested. Specified by design characterization.
3. It is recommended that TMS be asserted while TRST# is being deasserted.
Notes:
1. Platform support for VID transitions is required for the processor to operate within specifications.
Note: For Figure 2-11 through Figure 2-29, the following apply:
1. All System Reference Clock signal AC specifications are referenced to the Crossing
Voltage (VCROSS) of the BCLK_DP and BCLK_DN at rising edge of BCLK_DP.
2. All TAP signal group AC specifications are referenced to the TCK at 0.5 * VTT at the
processor lands. All TAP signal group timings (TMS, TDI, and so forth) are
referenced at 0.5 * VTT at the processor die (pads).
3. All CMOS signal AC specifications are referenced at 0.5 * VTT at the processor
lands.
The Intel QuickPath Interconnect electrical test setup are shown in Figure 2-11 and
Figure 2-12.
Figure 2-11. Intel QuickPath Interconnect Electrical Test Setup for Validating
Standalone TX Voltage and Timing Parameters
Ideal Loads
Silicon TX
Tx Package
Figure 2-12. Intel QuickPath Interconnect Electrical Test Setup for Validating
TX + Worst-Case Interconnect Specifications
W o r s t - C a s e In te r c o n n e c t Id e a l
Loads
S ilic o n
T x b it
( D a ta )
Tx
Package
S ilic o n Id e a l
T x b it Loads
( C lo c k )
L o s s le s s In te r c o n n e c t P h a s e
M a tc h e d to D a ta B it I n te r c o n n e c t
P-P n UI ac jitter
with Prob = 1E-7
P-P n UI ac jitter
with Prob = 1E-9
Voltage Margin
Distribution
Probability
= 1E -9
Timing Margin
Distribution
VRx-diff-pp-pin
Probability
= 1E -9
TRx-diff-pp-pin
650
600
550
Crossing Point (mV) 550 mV
500
550 + 0.5 (VHavg - 700)
450
400
250 + 0.5 (VHavg - 700)
350
300
250 mV
250
200
660 670 680 690 700 710 720 730 740 750 760 770 780 790 800 810 820 830 840 850
VHavg (mV)
Figure 2-17. Differential Clock Measurement Points for Duty Cycle and Period
Figure 2-18. Differential Clock Measurement Points for Rise and Fall time
Figure 2-19. Single-Ended Clock Measurement Points for Absolute Cross Point and Swing
Figure 2-20. Single-Ended Clock Measurement Points for Delta Cross Point
CK# (IMC)
CK(IMC)
BIOS
Programmable
Delay
MA, BS,
RAS#,
CAS#, WE# Tcmd_co Tcmd_co
(IMC)
Tcmd_cs Tcmd_cs
Control Signals
(IMC)
Note: Please refer to Table 2-18 for TAP Signal Group DC specifications and Table 2-27 for TAP Signal Group
AC specifications.
Figure 2-26. Test Reset (TRST#), Asynch GTL Input, and PROCHOT# Timing Waveform
Tq
Tr
THERMTRIP#
VCC
VTT
Tb
VTT VTT must be stable before VCCPWRGOOD
assertion
VCCPLL Ti
VDDQ
Te
Tm
VTTPWRGOOD
Figure 2-28. Voltage Sequence Timing Requirements
Tf VCCBOOT
VLFM
VCC
Th Varies based on BIOS execution
BCLK
Tj
Tk
RESET#
Electrical Specifications
Note: In order In order to ensure Timestamp Counter (TSC) synchronization across sockets in multi-socket
systems, the RESET# deassertion edge should arrive at the same BCLK rising edge at both sockets and
should meet the Tsu and Th requirement of 600ps relative to BCLK, as outlined in Table 2-26.
Note: This waveform illustrates an example of an Intel Adaptive Thermal Monitor transition or an Intel
Enhanced SpeedStep Technology transition that is six VID step down from the current state and six
steps back up. Any arbitrary up or down transition can be generalized from this waveform.
Data transfer requires the clean reception of data and clock signals. Ringing below
receiver thresholds, non-monotonic signal edges, and excessive voltage swings will
adversely affect system timings. Ringback and signal non-monotonicity cannot be
tolerated since these phenomena may inadvertently advance receiver state machines.
Excessive signal swings (overshoot and undershoot) are detrimental to silicon gate
oxide integrity, and can cause device failure if absolute voltage limits are exceeded.
Overshoot and undershoot can also cause timing degradation due to the build up of
inter-symbol interference (ISI) effects.
For these reasons, it is crucial that the designer work towards a solution that provides
acceptable signal quality across all systematic variations encountered in volume
manufacturing.
This section documents signal quality metrics used to derive topology and routing
guidelines through simulation. All specifications are specified at the processor die (pad
measurements).
Specifications for signal quality are for measurements at the processor core only and
are only observable through simulation. Therefore, proper signal simulation is the only
means of properly verifying timing and signal quality requirements.
The pulse magnitude, duration, and activity factor must all be used to determine if the
overshoot/undershoot pulse is within specifications.
Note: Oscillations below the reference voltage cannot be subtracted from the total
overshoot/undershoot pulse duration.
Notes:
1. These specifications are measured at the processor pin.
2. Refer to Figure 3-1 for description of Overshoot/Undershoot magnitude and duration.
Overshoot
Overshoot
Duration
Undershoot
Duration
Vss
Undershoot
4 Package Mechanical
Specifications
Die TIM
IHS
Substrate
Capacitors
LGA1366 Socket
LGA
System Board
Note:
1. Socket and baseboard are included for reference and are not part of processor package.
Dynamic Compressive Load 1779 N [400 lbf] [max static compressive + dynamic load] 1, 3, 4
Notes:
1. These specifications apply to uniform compressive loading in a direction normal to the processor IHS.
2. This is the maximum static force that can be applied by the heatsink and Independent Loading Mechanism
(ILM).
3. These specifications are based on limited testing for design characterization. Loading limits are for the
package constrained by the limits of the processor socket.
4. Dynamic loading is defined as an 11 ms duration average load superimposed on the static load
requirement.
5. See Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for minimum socket load
to engage processor within socket.
Shear 70 lbs
Tensile 25 lbs
Torque 35 in.lbs
5 Land Listing
Table 5-1. Land Name (Sheet 3 of 36) Table 5-1. Land Name (Sheet 4 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 5 of 36) Table 5-1. Land Name (Sheet 6 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 7 of 36) Table 5-1. Land Name (Sheet 8 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 9 of 36) Table 5-1. Land Name (Sheet 10 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 11 of 36) Table 5-1. Land Name (Sheet 12 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 13 of 36) Table 5-1. Land Name (Sheet 14 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 15 of 36) Table 5-1. Land Name (Sheet 16 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 17 of 36) Table 5-1. Land Name (Sheet 18 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 19 of 36) Table 5-1. Land Name (Sheet 20 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 21 of 36) Table 5-1. Land Name (Sheet 22 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 23 of 36) Table 5-1. Land Name (Sheet 24 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 25 of 36) Table 5-1. Land Name (Sheet 26 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 27 of 36) Table 5-1. Land Name (Sheet 28 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 29 of 36) Table 5-1. Land Name (Sheet 30 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 31 of 36) Table 5-1. Land Name (Sheet 32 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-1. Land Name (Sheet 33 of 36) Table 5-1. Land Name (Sheet 34 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
VSS V5 GND
VSS W3 GND
VSS W8 GND
VSS Y1 GND
VSS Y6 GND
Table 5-1. Land Name (Sheet 35 of 36) Table 5-1. Land Name (Sheet 36 of 36)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 3 of 35) Table 5-2. Land Number (Sheet 4 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 5 of 35) Table 5-2. Land Number (Sheet 6 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 7 of 35) Table 5-2. Land Number (Sheet 8 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 9 of 35) Table 5-2. Land Number (Sheet 10 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 11 of 35) Table 5-2. Land Number (Sheet 12 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 13 of 35) Table 5-2. Land Number (Sheet 14 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 15 of 35) Table 5-2. Land Number (Sheet 16 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 17 of 35) Table 5-2. Land Number (Sheet 18 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 19 of 35) Table 5-2. Land Number (Sheet 20 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 21 of 35) Table 5-2. Land Number (Sheet 22 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 23 of 35) Table 5-2. Land Number (Sheet 24 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 25 of 35) Table 5-2. Land Number (Sheet 26 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 27 of 35) Table 5-2. Land Number (Sheet 28 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 29 of 35) Table 5-2. Land Number (Sheet 30 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 31 of 35) Table 5-2. Land Number (Sheet 32 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
Table 5-2. Land Number (Sheet 33 of 35) Table 5-2. Land Number (Sheet 34 of 35)
Land Buffer Land Buffer
Land Name Direction Land Name Direction
No. Type No. Type
VSS Y6 GND
DDR_COMP[1] Y7 Analog
6 Signal Definitions
BPM#[7:0] I/O BPM#[7:0] are breakpoint and performance monitor signals. They are outputs
from the processor which indicate the status of breakpoints and programmable
counters used for monitoring processor performance. BPM#[7:0] should be
connected in a wired OR topology between all packages on a platform.
BPM#[5] and BPM#[7] signals between the two processors must remain
connected on production units.
CAT_ERR# I/O Indicates that the system has experienced a catastrophic error and cannot
continue to operate. The processor will set this for non-recoverable machine
check errors and other internal unrecoverable error. It is expected that every
processor in the system will have this hooked up in a wired-OR configuration.
Since this is an I/O pin, external agents are allowed to assert this pin which will
cause the processor to take a machine check exception.
On Intel Xeon processor 5600 series, CAT_ERR# is used for signalling the
following types of errors:
• Legacy MCERR’s, CAT_ERR# is pulsed for 16 BCLKs.
• Legacy IERR’s, CAT_ERR remains asserted until warm or cold reset.
QPI0_CLKRX_DN I Intel QuickPath Interconnect received clock is the input clock that corresponds
QPI0_CLKRX_DP to Intel QuickPath Interconnect port0 received data.
I
QPI0_CLKTX_DN O Intel QuickPath Interconnect forwarded clock sent with Intel QuickPath
QPI0_CLKTX_DP O Interconnect 0 port outbound data.
QPI1_CLKRX_DN I Intel QuickPath Interconnect received clock is the input clock that corresponds
QPI1_CLKRX_DP I to Intel QuickPath Interconnect 1 port received data.
QPI1_CLKTX_DN O Intel QuickPath Interconnect forwarded clock sent with Intel QuickPath
QPI1_CLKTX_DP O Interconnect port1 outbound data.
DBR# I DBR# is used only in systems where no debug port is implemented on the
system board. DBR# is used by a debug port interposer so that an in-target
probe can drive system reset.
DDR_THERM# I DDR_THERM# is used for imposing duty cycle throttling on all memory
channels. The platform should ensure that DDR_THERM# is asserted when any
DIMM is over T64.
DDR_THERM2# I DDR_THERM2# is used for imposing duty cycle throttling on all memory
channels or implementing 2X Refresh.
DDR{0/1/2}_BA[2:0] O Defines the bank which is the destination for the current Activate, Read, Write, 1
or Precharge command.
DDR{0/1/2}_CLK_N[3:0] O Differential clocks to the DIMM. All command and control signals are valid on
DDR{0/1/2}_CLK_P[3:0] the rising edge of clock.
DDR{0/1/2}_CS[7:0]# O Each signal selects one rank as the target of the command and address.
DDR{0/1/2}_DQS_N[17:0] I/O Differential pair, Data/ECC Strobe. Differential strobes latch data/ECC for each
DDR{0/1/2}_DQS_P[17:0] DRAM. Different numbers of strobes are used depending on whether the
connected DRAMs are x4,x8. Driven with edges in center of data, receive edges
are aligned with data edges.
DDR{0/1/2}_ECC[7:0] I/O Check Bits - An Error Correction Code is driven along with data on these lines
for DIMMs that support that capability.
DDR{0/1/2}_MA[15:0] O Selects the Row address for Reads and writes, and the column address for
activates. Also used to set values for DRAM configuration registers.
DDR{0/1/2}_ODT[3:0] O Enables various combinations of termination resistance in the target and non-
target DIMMs when data is read or written
DDR{0/1/2}_PAR_ERR#[2: I Parity Error detected by Registered DIMM (one for each DIMM).
0]
DDR{0/1/2}_RESET# O Resets DRAMs. Held low on power up, held high during self refresh, otherwise
controlled by configuration register.
ISENSE I Analog input voltage with respect to VSS for sensing core current consumption,
comes from VR11.1.
PECI I/O PECI (Platform Environment Control Interface) is the serial sideband interface
to the processor and is used primarily for thermal, power and error
management.
PECI_ID# I PECI_ID# is the PECI client address identifier. This pin is active low and
asserted when tied to VSS. Assertion of this pin results in a PECI client address
of 0x31 (versus the default 0x30 client address). This pin is primarily useful for
PECI client address differentiation in DP platforms and must be pulled up to
VTT on one socket and down to VSS on the other.
PREQ# I/O PREQ# is used by debug tools to request debug operation of the processor.
PROCHOT# I/O PROCHOT# will go active when the processor temperature monitoring sensor
detects that the processor has reached its maximum safe operating
temperature. This indicates that the processor Thermal Control Circuit has
been activated, if enabled. This signal can also be driven to the processor to
activate the Thermal Control Circuit.
If PROCHOT# is asserted at the deassertion of RESET# on DP enabled
products, the processor will tri-state its outputs. This signal does not have on-
die termination and must be terminated on the system board.
PSI# O Processor Power Status Indicator signal. This signal is asserted when maximum
possible processor core current consumption is less than 20A, Assertion of this
signal is an indication that the VR controller does not currently need to be able
to provide ICC above 20A, and the VR controller can use this information to
move to more efficient operation point.
RESET# I Asserting the RESET# signal resets the processor to a known state and
invalidates its internal caches without writing back any of their contents. Note
some PLL, Intel QuickPath Interconnect and error states are not affected by
reset and only VCCPWRGOOD forces them to a known state. For a power-on
Reset, RESET# must stay active for at least one millisecond after VCC and
BCLK have reached their proper specifications. RESET# must not be kept
asserted for more than 10 ms while VCCPWRGOOD is asserted. RESET# must
be held deasserted for at least one millisecond before it is asserted again.
RESET# must be held asserted before VCCPWRGOOD is asserted. This signal
does not have on-die termination and must be terminated on the system
board. RESET# is a common clock signal.
SKTOCC# O Socket occupied. The platform designer can use this signal to enable power
supplies when there is a CPU occupying the socket.
Requires external pull-up.
TAPPWRGOOD O Processor output signal, which when deasserted indicates the processor is in a
low power state and TAP functionality is unavailable.
TCK I TCK (Test Clock) provides the clock input for the processor Test Bus (also
known as the Test Access Port).
TDI I TDI (Test Data In) transfers serial test data into the processor. TDI provides the
serial input needed for JTAG specification support.
TDO O TDO (Test Data Out) transfers serial test data out of the processor. TDO
provides the serial output needed for JTAG specification support.
TMS I TMS (Test Mode Select) is a JTAG specification support signal used by debug
tools.
TRST# I TRST# (Test Reset) resets the Test Access Port (TAP) logic. TRST# must be
driven low during power on Reset.
VCCPWRGOOD I VCCPWRGOOD (Power Good) is a processor input. The processor requires this
signal to be a clean indication that BCLK, VCC, VCCPLL, VTTA and VTTD supplies
are stable and within their specifications. 'Clean' implies that the signal will
remain low (capable of sinking leakage current), without glitches, from the
time that the power supplies are turned on until they come within specification.
The signal must then transition monotonically to a high state. VCCPWRGOOD
can be driven inactive at any time, but BCLK and power must again be stable
before a subsequent rising edge of VCCPWRGOOD. In addition at the time
VCCPWRGOOD is asserted RESET# must be active. The VCCPWRGOOD signal
must be supplied to the processor; it is used to protect internal circuits against
voltage sequencing issues. It should be driven high throughout boundary scan
operation.
VTTA I Power for the analog portion of the Intel QuickPath and Shared Cache.
VTTD I Power for the digital portion of the Intel QuickPath and Shared Cache.
VDDPWRGOOD I VDDPWRGOOD is an input that indicates the VDDQ power supply is good. The
processor requires this signal to be a clean indication that the Vddq power
supply is stable and within their specifications. "Clean" implies that the signal
will remain low (capable of sinking leakage current), without glitches, from the
time that the VDDQ supply is turned on until it comes within specification. The
signals must then transition monotonically to a high state.
The VDDPWRGOOD signal must be supplied to the processor. This signal is
used to protect internal circuits against voltage sequencing issues.
2
VID[7:0] I/O VID[7:0] (Voltage ID) are output signals that are used to support automatic
selection of power supply voltages (VCC). The voltage supply for these signals
must be valid before the VR can supply VCC to the processor. Conversely, the
VR output must be disabled until the voltage supply for the VID signals become
valid. The VID signals are needed to support the processor voltage specification
variations. The VR must supply the voltage that is requested by the signals, or
disable itself.
MSID[2:0] - Market Segment ID, or MSID are provided to indicate the Market
Segment for the processor and may be used for future processor compatibility
or for keying. In addition, MSID protects the platform by preventing a higher
power processor from booting in a platform designed for lower power
processors. This value is latched from the platform in to the CPU, on the rising
edge of VTTPWRGOOD, during the cold boot power up sequence.
CSC[2:0] - Current Sense Configuration bits are output signals for ISENSE gain
setting.This value is latched on the rising edge of VTTPWRGOOD.
VTTPWRGOOD I The processor requires this input signal to be a clean indication that the VTT
power supply is stable and within their specifications. 'Clean' implies that the
signal will remain low (capable of sinking leakage current), without glitches,
from the time that the power supplies are turned on until they come within
specification. The signal must then transition monotonically to a high state. to
determine that the VTT voltage is stable and within specification. Note it is not
valid for VTTPWRGOOD to be deasserted while VCCPWRGOOD is asserted.
Notes:
1. DDR{0/1/2} refers to DDR3 Channel 0, DDR3 Channel 1, and DDR3 Channel 2.
2. VID[7:0] is an Input only during Power On Configuration. It is an Output signal during normal operation.
7 Thermal Specifications
A complete solution includes both component and system level thermal management
features. Component level thermal solutions can include active or passive heatsinks
attached to the processor integrated heat spreader (IHS). Typical system level thermal
solutions may consist of system fans combined with ducting and venting.
This section provides data necessary for developing a complete thermal solution. For
more information on designing a component level thermal solution, refer to the Intel®
Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide.
Note: The boxed processor will ship with a component thermal solution. Refer to Section 9 for
details on the boxed processor.
Thermal Profiles are broken out separately for the Intel Xeon processor 5600 series 6-
core and 4-core SKUs. This reflects different Thermal Test Vehicle (TTV) Correction
Factors (CFs), resulting from different power density characteristics associated with the
different number of cores. There is no difference in platform thermal solution
assumptions or boundary conditions between the 6-core and 4-core SKUs for a given
FMB (e.g., 130W). For the latest TTV CFs, please refer to the Intel® Xeon® Processor
5500/5600 Series Thermal/Mechanical Design Guide.
The upper point of the thermal profile consists of the Thermal Design Power (TDP) and
the associated TCASE value. It should be noted that the upper point associated with
Thermal Profile B (x = TDP and y = TCASE_MAX_B @ TDP) represents a thermal solution
design point. In actuality the processor case temperature will not reach this value due
to TCC activation.
Intel recommends that complete thermal solution designs target the Thermal Design
Power (TDP). The Adaptive Thermal Monitor feature is intended to help protect the
processor in the event that an application exceeds the TDP recommendation for a
sustained time period. To ensure maximum flexibility for future requirements, systems
should be designed to the Flexible Motherboard (FMB) guidelines, even if a processor
with lower power dissipation is currently planned. The Adaptive Thermal Monitor
feature must be enabled for the processor to remain within its specifications.
Notes:
1. These values are specified at VCC_MAX for all processor frequencies. Systems must be designed to ensure
the processor is not to be subjected to any static VCC and ICC combination wherein VCC exceeds VCC_MAX at
specified ICC. Please refer to the electrical loadline specifications in Section 2.
2. Thermal Design Power (TDP) should be used for processor thermal solution design targets. TDP is not the
maximum power that the processor can dissipate. TDP is measured at maximum TCASE.
3. Power specifications are defined at all VIDs found in Table 2-2. The processor may be delivered under
multiple VIDs for each frequency.
4. FMB, or Flexible Motherboard, guidelines provide a design target for meeting all planned processor
frequency requirements.
85
80
75
70
Y = 0.176 *x + 55.6
Temperature [C]
65
60
55
50
45
40
0 10 20 30 40 50 60 70 80 90 100 110 120 130
Power [W]
Notes:
1. The Frequency Optimized Server/Workstation Thermal Profile is representative of a volumetrically
unconstrained platform. Please refer to Table 7-2 for discrete points that constitute the thermal profile.
2. Implementation of the Frequency Optimized Server/Workstation Processor Thermal Profile should result in
virtually no TCC activation. Furthermore, utilization of thermal solutions that do not meet the thermal
profile will result in increased probability of TCC activation and may incur measurable performance loss.
3. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 55.6
10 57.4
20 59.1
30 60.9
40 62.7
50 64.4
60 66.2
70 67.9
80 69.7
90 71.5
100 73.2
110 75.0
120 76.8
130 78.5
85
80
75
70
Y = 0.190 *x + 55.7
Temperature [C]
65
60
55
50
45
40
0 10 20 30 40 50 60 70 80 90 100 110 120 130
Power [W]
Notes:
1. The Frequency Optimized Server/Workstation Thermal Profile is representative of a volumetrically
unconstrained platform. Please refer to Table 7-3 for discrete points that constitute the thermal profile.
2. Implementation of the Frequency Optimized Server/Workstation Processor Thermal Profile should result in
virtually no TCC activation. Furthermore, utilization of thermal solutions that do not meet the thermal
profile will result in increased probability of TCC activation and may incur measurable performance loss.
3. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 55.7
10 57.6
20 59.5
30 61.4
40 63.3
50 65.2
60 67.1
70 69.0
80 70.9
90 72.8
100 74.7
110 76.6
120 78.5
130 80.4
Notes:
1. These values are specified at VCC_MAX for all processor frequencies. Systems must be designed to ensure
the processor is not to be subjected to any static VCC and ICC combination wherein VCC exceeds VCC_MAX at
specified ICC. Please refer to the electrical loadline specifications in Section 2.
2. Thermal Design Power (TDP) should be used for processor thermal solution design targets. TDP is not the
maximum power that the processor can dissipate. TDP is measured at maximum TCASE.
3. Power specifications are defined at all VIDs found in Table 2-2. Processors may be delivered under multiple
VIDs for each frequency.
4. FMB, or Flexible Motherboard, guidelines provide a design target for meeting all planned processor
frequency requirements.
85
80
75
70
Thermal Profile B
Temperature [C]
Y = 0.257 * x + 56.9
65 Thermal Profile A
Y = 0.180 * x + 56.9
60
55
50
45
40
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95
Po we r [W]
Notes:
1. Thermal Profile A (refer to Table 7-5) is representative of a volumetrically unconstrained platform. Thermal
Profile B (refer to Table 7-6) is representative of a volumetrically constrained platform.
2. Implementation of Thermal Profile A should result in virtually no TCC activation. Utilization of thermal
solutions that do not meet Thermal Profile A will result in increased probability of TCC activation and may
incur measurable performance loss.
3. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 56.9
10 58.7
20 60.5
30 62.3
40 64.1
50 65.9
60 67.7
70 69.5
80 71.3
90 73.1
95 74.0
0 56.9
10 59.5
20 62.0
30 64.6
40 67.2
50 69.8
60 72.3
70 74.9
80 77.5
90 80.0
95 81.3
90
85
80
75
Thermal Profile B
Temperature [C]
70 Y = 0.272 * x + 57.1
Thermal Profile A
65
Y = 0.195 * x + 57.1
60
55
50
45
40
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95
Pow er [W]
Notes:
1. Thermal Profile A (refer to Table 7-7) is representative of a volumetrically unconstrained platform. Thermal
Profile B (refer to Table 7-8) is representative of a volumetrically constrained platform.
2. Implementation of Thermal Profile A should result in virtually no TCC activation. Utilization of thermal
solutions that do not meet Thermal Profile A will result in increased probability of TCC activation and may
incur measurable performance loss.
3. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 57.1
10 59.1
20 61.0
30 63.0
40 64.9
50 66.9
60 68.8
70 70.8
80 72.7
90 74.7
95 75.7
0 57.1
10 59.8
20 62.5
30 65.3
40 68.0
50 70.7
60 73.4
70 76.1
80 78.8
90 81.6
95 82.9
Notes:
1. These values are specified at VCC_MAX for all processor frequencies. Systems must be designed to ensure
the processor is not to be subjected to any static VCC and ICC combination wherein VCC exceeds VCC_MAX at
specified ICC. Please refer to the loadline specifications in Section 2.
2. Thermal Design Power (TDP) should be used for processor thermal solution design targets. TDP is not the
maximum power that the processor can dissipate. TDP is measured at maximum TCASE.
3. Power specifications are defined at all VIDs found in Table 2-2. The processor may be delivered under
multiple VIDs for each frequency.
4. FMB, or Flexible Motherboard, guidelines provide a design target for meeting all planned processor
frequency requirements.
80
75
70
Y = 0.306*x + 51.7
Temperature [C]
65
60
55
50
45
0 10 20 30 40 50 60 70 80
Powe r [W]
Notes:
1. The Thermal Profile is representative of a volumetrically constrained platform. Please refer to Table 7-10 for
discrete points that constitute the thermal profile.
2. Implementation of the Thermal Profile should result in virtually no TCC activation. Utilization of thermal
solutions that do not meet the Thermal Profile will result in increased probability of TCC activation and may
incur measurable performance loss.
3. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 51.7
10 54.8
20 57.8
30 60.9
40 63.9
50 67.0
60 70.1
70 73.1
80 76.2
80
75
70
Y = 0.321*x + 51.9
Temperature [C]
65
60
55
50
45
0 10 20 30 40 50 60 70 80
Power [W]
Notes:
1. The Thermal Profile is representative of a volumetrically constrained platform. Please refer to Table 7-11 for
discrete points that constitute the thermal profile.
2. Implementation of the Thermal Profile should result in virtually no TCC activation. Utilization of thermal
solutions that do not meet the Thermal Profile will result in increased probability of TCC activation and may
incur measurable performance loss.
3. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 51.9
10 55.1
20 58.3
30 61.5
40 64.8
50 68.0
60 71.2
70 74.4
80 77.6
Notes:
1. These values are specified at VCC_MAX for all processor frequencies. Systems must be designed to ensure
the processor is not to be subjected to any static VCC and ICC combination wherein VCC exceeds VCC_MAX at
specified ICC. Please refer to the loadline specifications in Section 2.
2. Thermal Design Power (TDP) should be used for processor thermal solution design targets. TDP is not the
maximum power that the processor can dissipate. TDP is measured at maximum TCASE.
3. Power specifications are defined at all VIDs found in Table 2-2. The processor may be shipped under
multiple VIDs for each frequency.
4. FMB, or Flexible Motherboard, guidelines provide a design target for meeting all planned processor
frequency requirements.
75
70
65
Y = 0.306*x + 51.0
Temperature [C]
60
55
50
45
0 10 20 30 40 50 60
Power [W]
Notes:
1. The Thermal Profile is representative of a volumetrically constrained platform. Please refer to Table 7-13 for
discrete points that constitute the thermal profile.
2. Implementation of the Thermal Profile should result in virtually no TCC activation. Utilization of thermal
solutions that do not meet the Thermal Profile will result in increased probability of TCC activation and may
incur measurable performance loss.
3. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
Table 7-13. Low Power Platform 60W Thermal Profile (6 Core) (Sheet 1 of 2)
Power (W) Maximum TCASE (°C)
0 51.0
10 54.0
20 57.1
Table 7-13. Low Power Platform 60W Thermal Profile (6 Core) (Sheet 2 of 2)
Power (W) Maximum TCASE (°C)
30 60.2
40 63.2
50 66.3
60 69.4
Notes:
1. These values are specified at VCC_MAX for all processor frequencies. Systems must be designed to ensure
the processor is not to be subjected to any static VCC and ICC combination wherein VCC exceeds VCC_MAX at
specified ICC. Please refer to the loadline specifications in Section 2.
2. Thermal Design Power (TDP) should be used for processor thermal solution design targets. TDP is not the
maximum power that the processor can dissipate. TDP is measured at maximum TCASE.
3. Power specifications are defined at all VIDs found in Table 2-2. The processor may be shipped under
multiple VIDs for each frequency.
4. FMB, or Flexible Motherboard, guidelines provide a design target for meeting all planned processor
frequency requirements.
65
60 Y = 0.318*x + 50.4
Temperature [C]
55
50
45
0 10 20 30 40
Power [W]
Notes:
1. The Thermal Profile is representative of a volumetrically constrained platform. Please refer to Table 7-15 for
discrete points that constitute the thermal profile.
2. Implementation of the Thermal Profile should result in virtually no TCC activation. Utilization of thermal
solutions that do not meet the Thermal Profile will result in increased probability of TCC activation and may
incur measurable performance loss.
3. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 50.4
10 53.6
20 56.8
30 59.9
40 63.1
Notes:
1. These values are specified at VCC_MAX for all processor frequencies. Systems must be designed to ensure
the processor is not to be subjected to any static VCC and ICC combination wherein VCC exceeds VCC_MAX at
specified ICC. Please refer to the loadline specifications in Section 2.
2. Thermal Design Power (TDP) should be used for processor thermal solution design targets. TDP is not the
maximum power that the processor can dissipate. TDP is measured at maximum TCASE.
3. Power specifications are defined at all VIDs found in Table 2-2. The processor may be shipped under
multiple VIDs for each frequency.
4. FMB, or Flexible Motherboard, guidelines provide a design target for meeting all planned processor
frequency requirements.
Notes:
1. The Thermal Profile is representative of a volumetrically constrained platform. Please refer to Table 7-17 for
discrete points that constitute the thermal profile.
2. Implementation of the nominal and short-term Thermal Profiles should result in virtually no TCC activation.
Utilization of thermal solutions that do not meet the Thermal Profile will result in increased probability of
TCC activation and may incur measurable performance loss.
3. The Nominal Thermal Profile must be used for all normal operating conditions, or for products that do not
require NEBS Level 3 compliance.
4. The Short-Term Thermal Profile may only be used for short-term excursions to higher ambient operating
temperatures, not to exceed 96 hours per instance, 360 hours per year, and a maximum of 15 instances
per year, as compliant with NEBS Level 3. Operation at the Short-Term Thermal Profile for durations
exceeding 360 hours per year violate the processor thermal specifications and may result in permanent
damage to the processor.
5. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 52 67
10 55 70
20 58 73
30 61 76
40 64 79
50 67 82
60 70 85
Notes:
1. These values are specified at VCC_MAX for all processor frequencies. Systems must be designed to ensure
the processor is not to be subjected to any static VCC and ICC combination wherein VCC exceeds VCC_MAX at
specified ICC. Please refer to the loadline specifications in Section 2.
2. Thermal Design Power (TDP) should be used for processor thermal solution design targets. TDP is not the
maximum power that the processor can dissipate. TDP is measured at maximum TCASE.
3. Power specifications are defined at all VIDs found in Table 2-2. The processor may be shipped under
multiple VIDs for each frequency.
4. FMB, or Flexible Motherboard, guidelines provide a design target for meeting all planned processor
frequency requirements.
Notes:
1. The Thermal Profile is representative of a volumetrically constrained platform. Please refer to Table 7-19 for
discrete points that constitute the thermal profile.
2. Implementation of the nominal and short-term Thermal Profiles should result in virtually no TCC activation.
Utilization of thermal solutions that do not meet the Thermal Profile will result in increased probability of
TCC activation and may incur measurable performance loss.
3. The Nominal Thermal Profile must be used for all normal operating conditions, or for products that do not
require NEBS Level 3 compliance.
4. The Short-Term Thermal Profile may only be used for short-term excursions to higher ambient operating
temperatures, not to exceed 96 hours per instance, 360 hours per year, and a maximum of 15 instances
per year, as compliant with NEBS Level 3. Operation at the Short-Term Thermal Profile for durations
exceeding 360 hours per year violate the processor thermal specifications and may result in permanent
damage to the processor.
5. Refer to the Intel® Xeon® Processor 5500/5600 Series Thermal/Mechanical Design Guide for system and
environmental implementation details.
0 50 65
5 53 68
10 56 71
15 58 73
20 61 76
25 64 79
30 67 82
35 69 84
40 72 87
Notes:
1. Figure is not to scale and is for reference only.
2. B1: Max = 45.07 mm, Min = 44.93 mm.
3. B2: Max = 42.57 mm, Min = 42.43 mm.
4. C1: Max = 39.1 mm, Min = 38.9 mm.
5. C2: Max = 36.6 mm, Min = 36.4 mm.
6. C3: Max = 2.3 mm, Min = 2.2 mm
7. C4: Max = 2.3 mm, Min = 2.2 mm.
Note: There is no specified correlation between DTS temperatures and processor case
temperatures; therefore it is not possible to use this feature to ensure the processor
case temperature meets the Thermal Profile specifications.
The Adaptive Thermal Monitor feature must be enabled for the processor to be
operating within specifications. The temperature at which Adaptive Thermal
Monitor activates the Thermal Control Circuit is not user configurable and is not
software visible. Snooping and interrupt processing are performed in the normal
manner while the TCC is active.
With a properly designed and characterized thermal solution, it is anticipated that the
TCC would only be activated for very short periods of time when running the most
power intensive applications. The processor performance impact due to these brief
periods of TCC activation is expected to be so minor that it would be immeasurable. An
under-designed thermal solution that is not able to prevent excessive activation of the
TCC in the anticipated ambient environment may cause a noticeable performance loss,
and in some cases may result in a TC that exceeds the specified maximum temperature
which may affect the long-term reliability of the processor. In addition, a thermal
solution that is significantly under-designed may not be capable of cooling the
processor even when the TCC is active continuously. Refer to the Intel® Xeon®
Processor 5500/5600 Series Thermal/Mechanical Design Guide for information on
designing a compliant thermal solution.
The duty cycle for the TCC, when activated by the Thermal Monitor, is factory
configured and cannot be modified. The Thermal Monitor does not require any
additional hardware, software drivers, or interrupt handling routines.
This method includes multiple operating points, each consisting of a specific operating
frequency and voltage. The first operating point represents the normal operating
condition for the processor. The remaining points consist of both lower operating
frequencies and voltages. When the TCC is activated, the processor automatically
transitions to the new operating frequency. This transition occurs very rapidly (on the
order of 2 µs).
Once the new operating frequency is engaged, the processor will transition to the new
core operating voltage by issuing a new VID code to the voltage regulator. The voltage
regulator must support dynamic VID steps to support this method. During the voltage
change, it will be necessary to transition through multiple VID codes to reach the target
operating voltage. Each step will be one VID table entry (see Table 2-2). The processor
continues to execute instructions during the voltage transition. Operation at the lower
voltages reduces the power consumption of the processor.
Temperature
fMAX
f1
f2
Frequency
VIDfMAX
VIDf1
VIDf2
VID
PROCHOT#
Time
The PROCHOT# signal is bi-directional in that it can either signal when the processor
(any core) has reached its maximum operating temperature or be driven from an
external source to activate the TCC. The ability to activate the TCC via PROCHOT# can
provide a means for thermal protection of system components.
PROCHOT# can allow voltage regulator (VR) thermal designs to target maximum
sustained current instead of maximum current. Systems should still provide proper
cooling for the VR, and rely on PROCHOT# only as a backup in case of system cooling
failure. The system thermal design should allow the power delivery circuitry to operate
within its temperature specification even while the processor is operating at its Thermal
Design Power.
The interface design was optimized for interfacing to Intel processor and chipset
components in both single processor and multiple processor environments. The single
wire interface provides low board routing overhead for the multiple load connections in
the congested routing area near the processor and chipset components. Bus speed,
error checking, and low protocol overhead provides adequate link bandwidth and
reliability to transfer critical device operating conditions and configuration information.
Intel recommends PECI host device speeds of 1.2 Mbps or lower for platforms using the
Intel Xeon processor 5600 series. PECI host devices operating at speeds greater than
1.2 Mbps may get a CPU “Timeout” error response to the PCI-ConfigRd() and
PCICconfigWr() PECI commands. This is expected to happen only during the deepest
idle states on Intel Xeon processor 5600 series. If higher bit rates are required,
platforms must be tolerant of “Timeout” completion codes during the deepest processor
package C-states. Please note that processors always request a 2 Mbps bit rate, and
will accept lower bit rates from a host according to timing negotiation specifications.
What follows is a processor-specific PECI client definition. PECI commands listed in
Table 7-20 apply to Intel Xeon processor 5600 series only.
Ping() Yes
GetDIB() Yes
GetTemp() Yes
PCIConfigRd() Yes
PCIConfigWr() Yes
1
MbxSend() Yes
1
MbxGet() Yes
Note:
1. Refer to Table 7-25 for a summary of mailbox commands supported by the Intel Xeon processor 5600
series.
PECI-based access to DRAM thermal readings and throttling control coefficients provide
a means for Board Management Controllers (BMCs) or other platform management
devices to feed hints into on-die memory controller throttling algorithms. These control
coefficients are accessible using PCI configuration space writes via PECI. The PECI-
based configuration write functionality is defined in Section 7.3.2.5, and the DRAM
throttling coefficient control functions are documented in the Intel® Xeon® Processor
5600 Series Datasheet, Volume 2.
The exact list of RAS-related registers in the PCI configuration space can be found in
the Intel® Xeon® Processor 5600 Series Datasheet, Volume 2.
Write Length: 0
Read Length: 0
Byte # 0 1 2 3
Byte # 0 1 2 3
Byte 0x30 0x00 0x00 0xe1
Definition
7.3.2.2 GetDIB()
The processor PECI client implementation of GetDIB() includes an 8-byte response and
provides information regarding client revision number and the number of supported
domains. All processor PECI clients support the GetDIB() command.
Write Length: 1
Read Length: 8
Command: 0xf7
Byte # 0 1 2 3 4
5 6 7 8 9
Revision
Device Info Reserved Reserved Reserved
Number
10 11 12 13
The Device Info byte gives details regarding the PECI client configuration. At a
minimum, all clients supporting GetDIB will return the number of domains inside the
package via this field. With any client, at least one domain (Domain 0) must exist.
Therefore, the Number of Domains reported is defined as the number of domains in
addition to Domain 0. For example, if the number 0b1 is returned, that would indicate
that the PECI client supports two domains.
7 6 5 4 3 2 1 0
Reserved
# of Domains
Reserved
All clients that support the GetDIB command also support Revision Number reporting.
The revision number may be used by a host or originator to manage different command
suites or response codes from the client. Revision Number is always reported in the
second byte of the GetDIB() response. The Revision Number always maps to the
revision number of this document.
7 4 3 0
Major Revision#
Minor Revision#
For a client that is designed to meet the PECI Specification, it returns ‘0010 0000b’.
7.3.2.3 GetTemp()
The GetTemp() command is used to retrieve the temperature from a target PECI
address. The temperature is used by the external thermal management system to
regulate the temperature on the die. The data is returned as a negative value
representing the number of degrees centigrade below the Thermal Control Circuit
Activation temperature of the PECI device. Note that a value of zero represents the
temperature at which the Thermal Control Circuit activates. The actual value that the
thermal management system uses as a control set point (Tcontrol) is also defined as a
negative number below the Thermal Control Circuit Activation temperature. TCONTROL
may be extracted from the processor by issuing a PECI Mailbox MbxGet() (see
Section 7.3.2.8), or using a RDMSR instruction.
Please refer to Section 7.3.6 for details regarding temperature data formatting.
Write Length: 1
Read Length: 2
Command: 0x01
Description: Returns the current temperature for addressed processor PECI client.
Byte # 0 1 2 3
4 5 6 7
Example bus transaction for a thermal sensor device located at address 0x30 returning
a value of negative 10° C:
Byte # 0 1 2 3
Byte 0x30 0x01 0x02 0x01
Definition
4 5 6 7
0xef 0x80 0xfd 0x4b
The typical client response is a passing FCS and good thermal data. Under some
conditions, the client’s response will indicate a failure.
General Sensor Error (GSE) Thermal scan did not complete in time. Retry is appropriate.
All other data Valid temperature reading, reported as a negative offset from the TCC
activation temperature.
7.3.2.4 PCIConfigRd()
The PCIConfigRd() command gives sideband read access to the entire PCI configuration
space maintained in the processor. This capability does not include support for route-
through to downstream devices or sibling processors. The exact listing of supported
devices, functions, and registers can be found in the Intel® Xeon® Processor 5600
Series Datasheet, Volume 2. PECI originators may conduct a device/function/register
enumeration sweep of this space by issuing reads in the same manner that BIOS
would. A response of all 1’s indicates that the device/function/register is
unimplemented.
PCI configuration addresses are constructed as shown in the following diagram. Under
normal in-band procedures, the Bus number (including any reserved bits) would be
used to direct a read or write to the proper device. Since there is a one-to-one mapping
between any given client address and the bus number, any request made with a non-
zero Bus number is ignored and the client will respond with a ‘pass’ completion code
but all 0’s in the data. The only legal bus number is 0x00. The client will return all 1’s in
the data response and ‘pass’ for the completion code for all of the following conditions:
• Unimplemented Device
• Unimplemented Function
• Unimplemented Register
31 28 27 20 19 15 14 12 11 0
Write Length: 5
Command: 0xc1
Description: Returns the data maintained in the PCI configuration space at the PCI
configuration address sent. The Read Length dictates the desired data return size. This
command supports byte, word, and dword responses as well as a completion code. All
command responses are prepended with a completion code that includes additional
pass/fail status information. Refer to Section 7.3.4.2 for details regarding completion
codes.
Byte # 0 1 2 3
4 5 6 7 8
9 10 8+RL 9+RL
Completion
Data 0 ... Data N FCS
Code
Note that the 4-byte PCI configuration address defined above is sent in standard PECI
ordering with LSB first and MSB last.
The typical client response is a passing FCS, a passing Completion Code (CC) and valid
Data. Under some conditions, the client’s response will indicate a failure.
CC: 0x80 Error causing a response timeout. Either due to a rare, internal timing condition or a processor
RESET or processor S1 state. Retry is appropriate outside of the RESET or S1 states.
7.3.2.5 PCIConfigWr()
The PCIConfigWr() command gives sideband write access to the PCI configuration
space maintained in the processor. The exact listing of supported devices, functions is
defined below in Table 7-23. PECI originators may conduct a device/function/register
enumeration sweep of this space by issuing reads in the same manner that BIOS
would.
Note:
1. Currently not available for access through the PECI PCIConfigWr() command.
PCI configuration addresses are constructed as shown in Figure 7-20, and this
command is subject to the same address configuration rules as defined in
Section 7.3.2.4. PCI configuration reads may be issued in byte, word, or dword
granularities.
Read Length: 1
Command: 0xc5
Description: Writes the data sent to the requested register address. Write Length
dictates the desired write granularity. The command always returns a completion code
indicating the pass/fail status information. Write commands issued to illegal (non-zero)
Bus Numbers, or unimplemented Device / Function / Register addresses are ignored
but return a passing completion code. Refer to Section 7.3.4.2 for details regarding
completion codes.
Byte # 0 1 2 3
4 5 6 7
8 WL-1
Completion
AW FCS FCS FCS
Code
Note that the 4-byte PCI configuration address and data defined above are sent in
standard PECI ordering with LSB first and MSB last.
The typical client response is a passing FCS, a passing Completion Code and valid Data.
Under some conditions, the client’s response will indicate a failure.
CC: 0x80 Error causing a response timeout. Either due to a rare, internal timing condition or a
processor RESET condition or processor S1 state. Retry is appropriate outside of the RESET
or S1 states.
7.3.2.6 Mailbox
The PECI mailbox (“Mbx”) is a generic interface to access a wide variety of internal
processor states. A Mailbox request consists of sending a 1-byte request type and
4-byte data to the processor, followed by a 4-byte read of the response data. The
following sections describe the Mailbox capabilities as well as the usage semantics for
the MbxSend and MbxGet commands which are used to send and receive data.
7.3.2.6.1 Capabilities
Ping 0x00 0x00 0x00 Verify the operability / existence of the Mailbox.
Thermal Status 0x01 Log bit clear Thermal Read the thermal status register and optionally clear any
Read/Clear mask Status Register log bits. The thermal status has status and log bits
indicating the state of processor TCC activation, external
PROCHOT# assertion, and Critical Temperature threshold
crossings.
Counter Clear 0x04 0x00 0x00 Concurrently clear and restart all counters.
Counter Read 0x05 Counter Counter Data Returns the counter number requested.
Number 0: Total reference time
1: Total TCC Activation time counter
Icc-TDC Read 0x06 0x00 Icc-TDC Returns the specified Icc-TDC of this part, in Amps.
Thermal Config 0x07 0x00 Thermal config Reads the thermal averaging constant.
Data Read data
Thermal Config 0x08 Thermal 0x00 Writes the thermal averaging constant.
Data Write Config Data
Tcontrol Read 0x09 0x00 Tcontrol Reads the fan speed control reference temperature,
Tcontrol, in PECI temperature format.
Machine Check 0x0A Bank Number Register Data Read processor Machine Check Banks.
Read / Index
T-State 0x0B 0x00 ACPI T-state Reads the PECI ACPI T-state throttling control word.
Throttling Control Word
Control Read
T-State 0x0C ACPI T-state 0x00 Writes the PECI ACPI T-state throttling control word.
Throttling Control Word
Control Write
Read Energy 0x0F 0x00 Energy Data Reads processor energy accumulator
Accumulator
Any MbxSend request with a request type not defined in Table 7-25 will result in a
failing completion code.
7.3.2.6.2 Ping
The Mailbox interface may be checked by issuing a Mailbox ‘Ping’ command. If the
command returns a passing completion code, it is functional. Under normal operating
conditions, the Mailbox Ping command should always pass.
The Thermal Status Read provides information on package level thermal status. Data
includes:
• The status of TCC activation
• Bidirectional PROCHOT# assertion
• Critical Temperature
These status bits are a subset of the bits defined in the IA32_THERM_STATUS MSR on
the processor, and more details on the meaning of these bits may be found in the
Intel 64 and IA-32 Architectures Software Developer’s Manual, Vol 3B.
Both status and sticky log bits are managed in this status word. All sticky log bits are
set upon a rising edge of the associated status bit, and the log bits are cleared only by
Thermal Status reads or a processor reset. A read of the Thermal Status Word always
includes a log bit clear mask that allows the host to clear any or all log bits that it is
interested in tracking.
A bit set to 0b0 in the log bit clear mask will result in clearing the associated log bit. If
a mask bit is set to 0b0 and that bit is not a legal mask, a failing completion code will
be returned. A bit set to 0b1 is ignored and results in no change to any sticky log bits.
For example, to clear the TCC Activation Log bit and retain all other log bits, the
Thermal Status Read should send a mask of 0xFFFFFFFD.
3
1 6 5 4 3 2 1 0
Reserved
A reference time and ‘Thermally Constrained’ time are managed in the processor. These
two counters are managed via the Mailbox. These counters are valuable for detecting
thermal runaway conditions where the TCC activation duty cycle reaches excessive
levels.
Total Time 0x00 Counts the total time the processor has been executing with a
resolution of approximately 1ms. This counter wraps at 32 bits.
Thermally Constrained Time 0x01 Counts the total time the processor has been operating at a
lowered performance due to TCC activation. This timer includes
the time required to ramp back up to the original P-state target
after TCC activation expires. This timer does not include TCC
activation time as a result of an external assertion of
PROCHOT#.
Icc-TDC is the processor Thermal Design Current specification. This data may be used
to confirm matching Icc profiles of processors in DP configurations. It may also be used
during the processor boot sequence to verify processor compatibility with baseboard
Icc delivery capabilities.
TCONTROL is used for fan speed control management. The TCONTROL limit may be
read over PECI using this Mailbox function. Unlike the in-band MSR interface, this
TCONTROL value is already adjusted to be in the native PECI temperature format of a
2-byte, 2’s complement number.
The Thermal Data Configuration register allows the PECI host to control the window
over which thermal data is filtered. The default window is 256 ms. The host may
configure this window by writing a Thermal Filtering Constant as a power of two.
E.g., sending a value of 9 results in a filtering window of 29 or 512 ms.
3
1 4 3 0
Reserved
PECI offers read access to all processor machine check banks. Bank numbers 2, 3, 4
and 5 are all associated with cores and therefore the appropriate core select must be
sent along with the Machine Check Read request. For all other Machine Check banks,
the Core Select must be set to 0x0.
It is possible that a fatal error may prevent access to other machine check banks. Host
controllers may read Power Control Unit errors directly by issuing a PCIConfigRd()
command of address 0x000000B0.
Byte # 0 1 2 3 4
Data 0x0A Bank Index Bank Number Core Select Reserved
Request Type Data[31:0]
0 0 MC0_CTL[31:0]
0 1 MC0_CTL[63:32]
0 2 MC0_STATUS[31:0]
0 3 MC0_STATUS[63:32]
0 4 MC0_ADDR[31:0]
0 5 MC0_ADDR[63:32]
0 6 MC0_MISC[31:0]
0 7 MC0_MISC[63:32]
1 0 MC1_CTL[31:0]
1 1 MC1_CTL[63:32]
1 2 MC1_STATUS[31:0]
1 3 MC1_STATUS[63:32]
1 4 MC1_ADDR[31:0]
1 5 MC1_ADDR[63:32]
1 6 MC1_MISC[31:0]
1 7 MC1_MISC[63:32]
2 0 MC2_CTL[31:0]
2 1 MC2_CTL[63:32]
2 2 MC2_STATUS[31:0]
2 3 MC2_STATUS[63:32]
2 4 MC2_ADDR[31:0]
2 5 MC2_ADDR[63:32]
2 6 MC2_MISC[31:0]
2 7 MC2_MISC[63:32]
3 0 MC3_CTL[31:0]
3 1 MC3_CTL[63:32]
3 2 MC3_STATUS[31:0]
3 3 MC3_STATUS[63:32]
3 4 MC3_ADDR[31:0]
3 5 MC3_ADDR[63:32]
3 6 MC3_MISC[31:0]
3 7 MC3_MISC[63:32]
4 0 MC4_CTL[31:0]
4 1 MC4_CTL[63:32]
4 2 MC4_STATUS[31:0]
4 3 MC4_STATUS[63:32]
4 4 MC4_ADDR[31:0]
4 5 MC4_ADDR[63:32]
4 6 MC4_MISC[31:0]
4 7 MC4_MISC[63:32]
5 0 MC5_CTL[31:0]
5 1 MC5_CTL[63:32]
5 2 MC5_STATUS[31:0]
5 3 MC5_STATUS[63:32]
5 4 MC5_ADDR[31:0]
5 5 MC5_ADDR[63:32]
0 4 MC0_ADDR[31:0]
0 5 MC0_ADDR[63:32]
0 6 MC0_MISC[31:0]
0 7 MC0_MISC[63:32]
1 0 MC1_CTL[31:0]
1 1 MC1_CTL[63:32]
1 2 MC1_STATUS[31:0]
1 3 MC1_STATUS[63:32]
1 4 MC1_ADDR[31:0]
1 5 MC1_ADDR[63:32]
1 6 MC1_MISC[31:0]
1 7 MC1_MISC[63:32]
2 0 MC2_CTL[31:0]
2 1 MC2_CTL[63:32]
2 2 MC2_STATUS[31:0]
2 3 MC2_STATUS[63:32]
2 4 MC2_ADDR[31:0]
2 5 MC2_ADDR[63:32]
2 6 MC2_MISC[31:0]
2 7 MC2_MISC[63:32]
3 0 MC3_CTL[31:0]
3 1 MC3_CTL[63:32]
3 2 MC3_STATUS[31:0]
3 3 MC3_STATUS[63:32]
3 4 MC3_ADDR[31:0]
3 5 MC3_ADDR[63:32]
3 6 MC3_MISC[31:0]
3 7 MC3_MISC[63:32]
4 0 MC4_CTL[31:0]
4 1 MC4_CTL[63:32]
4 2 MC4_STATUS[31:0]
4 3 MC4_STATUS[63:32]
4 4 MC4_ADDR[31:0]
4 5 MC4_ADDR[63:32]
4 6 MC4_MISC[31:0]
4 7 MC4_MISC[63:32]
5 0 MC5_CTL[31:0]
5 1 MC5_CTL[63:32]
5 2 MC5_STATUS[31:0]
5 3 MC5_STATUS[63:32]
5 4 MC5_ADDR[31:0]
5 5 MC5_ADDR[63:32]
5 6 MC5_MISC[31:0]
5 7 MC3_MISC[63:32]
6 0 MC6_CTL[31:0]
6 1 MC6_CTL[63:32]
6 2 MC6_STATUS[31:0]
6 3 MC6_STATUS[63:32]
6 4 MC6_ADDR[31:0]
6 5 MC6_ADDR[63:32]
6 6 MC6_MISC[31:0]
6 7 MC6_MISC[63:32]
8 0 MC8_CTL[31:0]
8 1 MC8_CTL[63:32]
8 2 MC8_STATUS[31:0]
8 3 MC8_STATUS[63:32]
8 4 MC8_ADDR[31:0]
8 5 MC8_ADDR[63:32]
8 6 MC8_MISC[31:0]
8 7 MC8_MISC[63:32]
PECI offers the ability to enable and configure ACPI T-state (core clock modulation)
throttling. ACPI T-state throttling forces all CPU cores into duty cycle clock modulation
where the core toggles between C0 (clocks on) and C1 (clocks off) states at the
specified duty cycle. This throttling reduces CPU performance to the duty cycle
specified and, more importantly, results in processor power reduction.
The processor supports software initiated T-state throttling and automatic T-state
throttling as part of the internal Thermal Monitor response mechanism (upon TCC
activation). The PECI T-state throttling control register read/write capability is
managed only in the PECI domain. In-band software may not manipulate or read the
PECI T-state control setting. In the event that multiple agents are requesting T-state
throttling simultaneously, the CPU always gives priority to the lowest power setting, or
the numerically lowest duty cycle.
The only supported duty cycle is 12.5% (12.5% clocks on, 87.5% clocks off). It is
expected that T-state throttling will be engaged only under emergency thermal or
power conditions. Future products may support more duty cycles, as defined in the
following table:
0x0 Undefined
Intel Xeon processor 5600 series energy consumption may be monitored with regular
reads of the free running Energy Accumulator. The Energy Accumulator reports
processor energy consumed on the VCC, VTT, VCCPLL and VDD power rails in Joules in
an unsigned, 32-bit fixed point binary format. The least-significant 16 bits report the
fractional data and the most significant 16 bits report the integer data. Because the
Energy Accumulator is constantly incrementing, the data is only relevant as a delta
between two subsequent reads.
Overflow in the accumulator is infrequent, but it will occur constantly during normal
execution. Therefore, the PECI host controller or firmware must be designed such that
it can guarantee a polling rate sufficient to prevent aliasing. Recommended polling
rates are between 1 Hz and 10 Hz. As an example, a processor steadily consuming
130 W of power will overflow the accumulator once every 8.4 minutes.
The free running nature of the PECI Energy Accumulator register allows for irregular
timing on reads. It is expected that the PECI host controller or firmware will manage its
own time stamps.
Time: 127 ms
EnergyAccumulated[0] = MbxGet(TransID)
EnergyAccumulated[0] = 0xfff737a3
Time: 228 ms
EnergyAccumulated[1] = MbxGet(TransID)
EnergyAccumulated[0] = 0x00176dc
else {
This capability may not be available on all products. To enumerate availability, users
may issue an Energy Accumulator Read command. Products that return a passing
completion code support the command.
7.3.2.7 MbxSend()
The MbxSend() command is utilized for sending requests to the generic Mailbox
interface. Those requests are in turn serviced by the processor with some nominal
latency and the result is deposited in the mailbox for reading. MbxGet() is used to
retrieve the response and details are documented in Section 7.3.2.8.
The details of processor mailbox capabilities are described in Section 7.3.2.6.1, and
many of the fundamental concepts of Mailbox ownership, release, and management are
discussed in Section 7.3.2.9.
Regardless of the function of the mailbox command, a request type modifier and 4-byte
data payload must be sent. For Mailbox commands where the 4-byte data field is not
applicable (e.g., the command is a read), the data written should be all zeroes.
Byte # 0 1 2 3 4
Byte Request Type Data[31:0]
Definition
Write Length: 7
Read Length: 1
Command: 0xd1
Description: Deposits the Request Type and associated 4-byte data in the Mailbox
interface and returns a completion code byte with the details of the execution results.
Refer to Section 7.3.4.2 for completion code definitions.
Byte # 0 1 2 3
4 5 6 7 8
9 10 11 12
Completion
AW FCS FCS FCS
Code
Note that the 4-byte data defined above is sent in standard PECI ordering with LSB first
and MSB last.
0x80 Error causing a response timeout. Either due to a rare, internal timing condition or a
processor RESET condition or processor S1 state. Retry is appropriate outside of the
RESET or S1 states.
If MbxSend() response returns a bad FCS in the data, the completion code can't be
trusted and the semaphore may or may not be taken. In order to clean out the
interface, an MbxGet() must be issued and the response data should be discarded.
7.3.2.8 MbxGet()
The MbxGet() command is utilized for retrieving response data from the generic
Mailbox interface as well as for unlocking the acquired mailbox. Please refer to
Section 7.3.2.7 for details regarding the MbxSend() command. Many of the
fundamental concepts of Mailbox ownership, release, and management are discussed
in Section 7.3.2.9.
Any mailbox request made with an illegal or unlocked Transaction ID will get a failed
completion code response. If the Transaction ID matches an outstanding transaction ID
associated with a locked mailbox, the command will complete successfully and the
response data will be returned to the originator.
Unlike MbxSend(), no Assured Write protocol is necessary for this command because
this is a read-only function.
Write Length: 2
Read Length: 5
Command: 0xd5
Description: Retrieves response data from mailbox and unlocks / releases that
mailbox resource.
Byte # 0 1 2 3
4 10
5 11
6
Completion
Transaction ID FCS
Code
7 10
5
8 11
6
9 10 11
Note that the 4-byte data response defined above is sent in standard PECI ordering
with LSB first and MSB last.
Aborted Write FCS Response data is not ready. Command retry is appropriate.
0x80 Error causing a response timeout. Either due to a rare, internal timing condition or a
processor RESET condition or processor S1 state. Retry is appropriate outside of the
RESET or S1 states.
0x88 Machine Check Banks is currently unavailable (selected core is asleep or unavailable)
The MbxSend() command is used to acquire control of the PECI mailbox and issue
information regarding the specific request. The completion code response indicates
whether or not the originator has acquired a lock on the mailbox, and that completion
code always specifies the Transaction ID associated with that lock (see
Section 7.3.2.9.2).
Once a mailbox has been acquired by an originating agent, future requests to acquire
that mailbox will be denied with an ‘interface busy’ completion code response.
The lock on a mailbox is not achieved until the last bit of the MbxSend() Read FCS is
transferred (in other words, it is not committed until the command completes). If the
host aborts the command at any time prior to that bit transmission, the mailbox lock
will be lost and it will remain available for any other agent to take control.
7.3.2.9.2 Transaction ID
For all MbxSend() commands that complete successfully, the passing completion code
(0x4X) includes a 4-bit Transaction ID (‘X’). That ID is the key to the mailbox and must
be sent when retrieving response data and releasing the lock by using the MbxGet()
command.
The mailbox is a shared resource that can result in artificial bandwidth conflicts among
multiple querying processes that are sharing the same originator interface. The
interface response time is quick, and with rare exception, back to back MbxSend() and
MbxGet() commands should result in successful execution of the request and release of
the mailbox. In order to guarantee timely retrieval of response data and mailbox
release, the mailbox semaphore has a timeout policy. If the PECI bus has a cumulative
‘0 time of 1ms since the semaphore was acquired, the semaphore is automatically
cleared. In the event that this timeout occurs, the originating agent will receive a failed
completion code upon issuing a MbxGet() command, or even worse, it may receive
corrupt data if this MbxGet() command so happens to be interleaved with an
MbxSend() from another process. Please refer to Table 7-30 for more information
regarding failed completion codes from MbxGet() commands.
Timeouts are undesirable, and the best way to avoid them and guarantee valid data is
for the originating agent to always issue MbxGet() commands immediately following
MbxSend() commands.
If the timeout policy is too restrictive, it can be disabled. BIOS may write
MSR_MISC_POWER_MGMT (0x1AA), bit 11 to 0b1 in order to force a disable of this
automatic timeout.
The PECI mailbox interface is designed to have response data available within plenty of
margin to allow for back-to-back MbxSend() and MbxGet() requests. However, under
rare circumstances that are out of the scope of this specification, it is possible that the
response data is not available when the MbxGet() command is issued. Under these
circumstances, the MbxGet() command will respond with an Abort FCS and the
originator should re-issue the MbxGet() request.
0b01 0
0b10 1
• Assured Write FCS (AW FCS) failure. Note that under most circumstances, an
Assured Write failure will appear as a bad FCS. However, when an originator issues
a poorly formatted command with a miscalculated AW FCS, the client will
intentionally abort the FCS in order to guarantee originator notification.
An originator that is decoding these commands can apply a simple mask to determine
pass or fail. Bit 7 is always set on a failed command, and is cleared on a passing
command.
0x80 Error causing a response timeout. Either due to a rare, internal timing condition or a
processor RESET condition or processor S1 state. Retry is appropriate outside of the RESET
or S1 states.
0x88 Machine Check Banks is currently unavailable (selected core is asleep or unavailable)
Note: The codes explicitly defined in this table may be useful in PECI originator response
algorithms. All reserved or undefined codes may be generated by a PECI client device,
and the originating agent must be capable of tolerating any code. The Pass/Fail mask
defined in Table 7-33 applies to all codes and general response policies may be based
on that limited information.
Please refer to the definition of each command in Section 7.3.2 for a specific definition
of possible command codes or FCS responses for a given command. The following
response policy definition is generic, and more advanced response policies may be
employed at the discretion of the originator developer.
Abort FCS Retry Fail with PECI client device error. May be due to illegal command codes.
Fail Retry Either the PECI client doesn’t support the current command code, or it has
failed in its attempts to construct a response.
None (all 0’s) Force bus idle Fail with PECI client device error. Client may be dead or otherwise non-
(1ms low), retry responsive (in RESET or S1, for example).
Pass Pass NA
7.3.6.2 Interpretation
The resolution of the processor’s Digital Thermal Sensor (DTS) is approximately 1°C,
which can be confirmed by a RDMSR from IA32_THERM_STATUS MSR (0x19C) where it
is architecturally defined. PECI temperatures are sent through a configurable low-pass
filter prior to delivery in the GetTemp() response data. The output of this filter produces
temperatures at the full 1/64°C resolution even though the DTS itself is not this
accurate.
Temperature readings from the processor are always negative, and imply an offset
from the reference TCC activation temperature. As an example, assume that the TCC
activation temperature reference is 100°C. A PECI thermal reading of -10°C indicates
that the processor is running at 10°C below the TCC activation temperature, or 90°C.
PECI temperature readings are not reliable at temperatures above TCC activation (since
the processor is operating out of spec at this temperature). Therefore, the readings are
never positive.
Please refer to Section 7.3.2.6 for the definition of the thermal configuration command.
Table 7-37. PECI Client Response During Power-Up (During ‘Data Not Ready’)
Command Response
MbxGet() Client responds with Abort FCS (if MbxSend() has been previously issued)
In the event that the processor is tri-stated using power-on-configuration controls, the
PECI client will also be tri-stated.
Vtt
VttPwrGd
SupplyVcc
Bclk
VccPwrGd
RESET#
Mclk
The client address may not be changed after VCCPWRGOOD assertion, until the next
power cycle on the processor. Removal of a processor from its socket or tri-stating a
processor in a DP configuration will have no impact to the remaining non-tri-stated
PECI client address.
7.3.7.4 C-States
The Intel Xeon processor 5600 series are fully functional under all core and package
C-states. Support for package C-states is a function of processor SKU and platform
capabilities. All package C-states (C1E, C3, C6) are documented here for completeness,
but actual processor support for these C-states may vary.
Because the processor takes aggressive power savings actions under the deepest
C-states, PECI requests may have an impact to platform power. The impact is
documented below:
7.3.7.5 S-States
The PECI client is always guaranteed to be operational under S0 and S1 sleep states.
Under S3 and deeper sleep states, the PECI client response is undefined and therefore
unreliable.
Table 7-38 specifies absolute maximum and minimum storage temperature limits which
represent the maximum and minimum device condition beyond which damage, latent
or otherwise, may occur. The table also specifies sustained storage temperature,
relative humidity, and time-duration limits. At conditions outside sustained limits, but
within absolution maximum and minimum ratings, quality & reliability may be affected.
Notes:
1. Storage conditions are applicable to storage environments only. In this scenario, the processor must not
receive a clock, and no lands can be connected to a voltage bias. Storage within these limits will not affect
the long-term reliability of the device. For functional operation, please refer to the processor case
temperature specifications.
2. These ratings apply to the Intel component and do not include the tray or packaging.
3. Failure to adhere to this specification can affect the long-term reliability of the processor.
4. Non operating storage limits post board attache: Storage conditions limits for the component once attached
to the application board are not specified.
8 Features
Please note that requests to execute Built-In Self Test (BIST) are not selected by
hardware, but rather passed across the Intel QuickPath Interconnect link during
initialization.
Notes:
1. Asserting the signal during RESET# de-assertion will select the corresponding option. Once selected, this
option cannot be changed except via another reset. The processor does not distinguish between a “warm”
reset and a “power-on” reset. Output tri-state via the PROCHOT# power-on configuration option is referred
to as Fault Resilient Boot (FRB).
2. Latched when VTTPWRGOOD is asserted and all internal power good conditions are met.
BCLK
CPURESET#
Tri-State POC
(xxPROCHOT#)
Non-FRB assertion of
xxPROCHOT# during this window
can trigger false tri-state
xxPROCHOT# deassertion is not required for FRB
Note:
1. FRB = Fault-Resilient Boot
Power-On Configuration (POC) logic levels are MUX’ed onto the VID[7:0] signals with
1-5 kΩ pull-up and pull-down resistors located on the baseboard. These include:
• VID[2:0] / MSID[2:0] = Market Segment ID
• VID[5:3] / CSC [2:0] = Current Sense Configuration
• VID[6] = Reserved
• VID[7] = VR11.1 Select
Pull-up and pull-down resistors on the baseboard eliminate the need for timing
specifications. After OUTEN signal is asserted, the VID[7:0] CMOS drivers (typically 50
Ω up/down impedance) over-ride the POC pull-up/down resistors located on the
baseboard and drive the necessary VID pattern. Please refer to Table 2-3 for further
details.
Note: Software may make C-state requests by using a legacy method involving I/O reads
from the ACPI-defined processor clock control registers, referred to as P_LVLx. This
feature is designed to provide legacy support for operating systems that initiate C-state
transitions via access to pre-defined ICH registers. The base P_LVLx register is P_LVL2,
corresponding to a C3 request. P_LVL3 is C6, and all P_LVL4+ are demoted to a C6.
P_LVLx is limited to a subset of C-states (For example, P_LVL8 is not supported and will
not cause an I/O redirection to a C8 request. Instead, it will fall through like a normal
I/O instruction). The range of I/O addresses that may be converted into C-state
requests is also defined in the PMG_IO_CAPTURE MSR, in the ‘C-state Range’ field. This
field maybe written by BIOS to restrict the range of I/O addresses that are trapped and
redirected to MWAIT instructions. Note that when I/O instructions are used, no MWAIT
substates can be defined, as therefore the request defaults to have a sub-state or zero,
but always assumes the ‘Break on EFLAGS.IF==0’ control that can be selected using
ECX with an MWAIT instruction.
C0
MWAIT C1,
HLT
2
2 2 2
MWAIT C1, MWAIT C6,
HLT MWAIT C3, I/O C6
(C1E enabled) I/O C3
1 1
C1 C1E C3 C6
C0 C0 C0 C0 C0
C3 C0 C11 C3 C3
1
C6 C0 C1 C3 C6
Note:
1. If enabled, state will be C1E.
8.2.1.1 C0 State
This is the normal operating state in the processor.
A System Management Interrupt (SMI) handler will return execution to either Normal
state or the C1 state. See the Intel® 64 and IA-32 Architecture Software Developer's
Manuals, Volume III: System Programmer's Guide for more information.
While in C1/C1E state, the processor will process bus snoops and snoops from the
other threads.
To operate within specification, BIOS must enable the C1E feature for all installed
processors.
8.2.1.3 C3 State
Individual threads of the processor can enter the C3 state by initiating a P_LVL2 I/O
read to the P_BLK or an MWAIT(C3) instruction. Before entering core C3, that core
flushes the contents of its caches. Caches shared among cores are not impacted.
Except for the caches, the processor core maintains all its architectural state while in
the C3 state. All of the clocks in the processor core are stopped in the C3 state.
Because the core’s caches are flushed, the processor keeps the core in the C3 state
when the processor detects a snoop on the Intel QuickPath Interconnect Link or when
another logical processor in the same package accesses cacheable memory. The
processor core will transition to the C0 state upon occurrence of an interrupt. RESET#
will cause the processor core to initialize itself.
8.2.1.4 C6 State
Individual threads of the processor can enter the C6 state by initiating a P_LVL3 read to
the P_BLK or an MWAIT(C6) instruction. Before entering core C6, that core flushes the
contents of its caches. Caches shared among cores are not impacted. The processor
achieves additional power savings in the core C6 state.
If Intel QuickPath Interconnect L1 has been granted, the processor will disable some
clocks and PLLs and for processors with an integrated memory controller, the DRAM will
be put into self-refresh.
If Intel QuickPath Interconnect L1 has been granted, the processor will disable some
clocks and PLLs. The shared cache will enter a deep sleep state. Additionally, for
processors with an integrated memory controller, the DRAM will be put into self-refresh.
C1E 37 W 32 W 37/40 W 25 W 22 W
C3 32 W 28 W 33/33 W 20 W 18 W
C6 12 W 10 W 12/14 W 8W 8W
Notes:
1. Specifications are at TCASE = 50 °C with all cores in the specified C-state.
2. Standard/Basic SKUs.
S1 Cores in C1E like state, processor responds with S0 (via reset or PMReq)
CmpD(S1) message. S3, S4 (via PMReq)
Notes:
1. If the chipset requests an S-state transition which is not allowed, a machine check error will be generated
by the processor.
9.1 Introduction
Intel boxed processors are intended for system integrators who build systems from
components available through distribution channels. The Intel Xeon processor 5600
series will be offered as an Intel boxed processor, however the thermal solution will be
sold separately.
Intel Xeon processor 5600 series boxed processors will not include a thermal solution in
the box. Intel will offer boxed thermal solutions separately through the same
distribution channels. Please reference Section 9.1.1 - Section 9.1.4 for more details on
Boxed Processor Thermal Solutions.
Figure 9-1. STS100C Passive / Active Combination Heat Sink (with Removable Fan)
Figure 9-2. STS100C Passive / Active Combination Heat Sink (with Fan Removed)
The STS100C and STS100A utilize a fan capable of 4-pin pulse width modulated (PWM)
control. Use of a 4-pin PWM controlled active thermal solution helps customers meet
acoustic targets in pedestal platforms through the baseboard’s ability to directly control
the RPM of the processor heat sink fan. See Section 9.3 for more details on fan speed
control. Also see Section 7.3 for more on the PWM and PECI interface along with Digital
Thermal Sensors (DTS).
None of the heat sink solutions exceed a mass of 550 grams. Note that this is per
processor, a dual processor system will have up to 1100 grams total mass in the heat
sinks.
90.00
[3.543 ]
MAX THERMAL
RETENTION OUTLINE
80.00
[3.150 ]
THERMAL RETENTION
HOLE PATTERN
61.20
[2.409 ]
SOCKET ILM
HOLE PATTERN
B 49.90
LEGEND, THIS SHEET ONLY B
[1.965 ]
SOCKET BODY OUTLINE, ZONE 1:
SOCKET BODY OUTLINE FOR REFERENCE ONLY 0.0 MM MAX COMPONENT HEIGHT, NO COMPONENT PLACEMENT,
FOR REFERENCE ONLY SOCKET, ILM, AND FINGER ACCESS KEEPIN ZONE
44.70
[1.760 ]
CENTERLINE OF OUTER ZONE 2:
SOCKET BALL ARRAY 7.0 MM MAX COMPONENT HEIGHT 7
ZONE 3:
3.0 MM MAX COMPONENT HEIGHT 7
ZONE 4:
LINE REPRESENTS OF
0.0 MM MAX COMPONENT HEIGHT, NO COMPONENT PLACEMENT
OUTERMOST ROWS AND COLUMNS
RETENTION MODULE OR HEAT SINK TOUCH ZONE
OF SOCKET BALL ARRAY OUTLINE.
FOR REFERENCE ONLY
ZONE 5:
0.0 MM MAX COMPONENT HEIGHT, NO COMPONENT PLACEMENT,
NO ROUTE ZONE
ZONE 6:
BALL 1 POSITION 4 1.97 MM MAX COMPONENT HEIGHT, SOCKET CAVITY 7
8 7 6 5 4 3 2 1
171
172
Figure 9-6.
D D
4X 6.00 +0.06
[ 0.295 ]
[ 2.854 ]
[ 3.150 ]
[0.000]
4X 3.80 NPTH
[ 0.236 ] -0.03
+0.002
5.00
[ 0.197 ]
2X 7.50
9.60
[ 0.378 ]
12.30
[ 0.484 ]
19.17
[ 0.755 ]
BALL 1 4
22.000
[ 0.8661 ]
32.85
[ 1.293 ]
47.15
[ 1.856 ]
58.000
[ 2.2835 ]
67.70
[ 2.665 ]
2X 72.50
2X 80.00
85.00
[ 3.346 ]
2X 0.00
NO ROUTE 0.150
COPPER PAD ON SURFACE [ -0.001 ]
SOCKET ILM
MOUNTING HOLES
5.00
[ 0.197 ]
2X 0.00
[0.000]
3.30
[ 0.130 ]
2X 7.50
[ 0.295 ]
2X 9.400 +0.06
4X 4.03 NPTH
[ 0.3701 ] -0.03
+0.002
C 0.159 C
9.900 [ -0.001 ]
[ 0.3898 ] THERMAL RETENTION
MOUNTING HOLES
29.90
[ 1.177 ]
30.600
[ 1.205 ]
4X 6.00
[ 0.236 ]
DETAIL A
SCALE 6.000
49.40
[ 1.945 ]
62.39
[ 2.456 ]
B BALL 1 4 B
2X 70.600
[ 2.7795 ]
Top Side Baseboard Mounting-Hole Keep-Out Zones
2X 72.50
[ 2.854 ]
77.90
LEGEND, THIS SHEET ONLY
[ 3.067 ]
ZONE 1:
3X 80.00 0.0 MM MAX COMPONENT HEIGHT, NO COMPONENT PLACEMENT,
SOCKET, ILM, AND FINGER ACCESS KEEPIN ZONE
[ 3.150 ]
85.00 ZONE 2:
[ 3.346 ] 7.0 MM MAX COMPONENT HEIGHT 7
ZONE 4:
0.0 MM MAX COMPONENT HEIGHT, NO COMPONENT PLACEMENT
RETENTION MODULE OR HEAT SINK TOUCH ZONE
ZONE 5:
A 0.0 MM MAX COMPONENT HEIGHT, NO COMPONENT PLACEMENT, A
NO ROUTE ZONE
ZONE 6:
AS VIEWED FROM PRIMARY SIDE 1.97 MM MAX COMPONENT HEIGHT, SOCKET CAVITY 7
OF THE MOTHERBOARD
DEPARTMENT SIZE DRAWING NUMBER REV
R
2200 MISSION COLLEGE BLVD.
(DETAILS) EASD / PTMI
P.O. BOX 58119 D D77712 02
SANTA CLARA, CA 95052-8119
SCALE: 3.000 DO NOT SCALE DRAWINGSHEET 2 OF 4
8 7 6 5 4 3 2 1
Boxed Processor Specifications
D D
85.00
[ 3.346 ]
70.50
[ 2.776 ]
47.15
[ 1.856 ]
32.85
[ 1.293 ]
9.50
[ 0.374 ]
0.00
[0.000]
5.00
[ 0.197 ]
5.00
[ 0.197 ]
0.00
[0.000]
5.00
[ 0.197 ]
DESKTOP BACKPLATE
KEEPIN SHOWN FOR
REFERENCE ONLY 17.17
C C
30.60
[ 1.205 ]
(90.00 ) (72.20 )
[3.543 ] [2.843 ]
49.40
[ 1.945 ]
Bottom Side Baseboard Keep-Out Zones
62.83
[ 2.474 ]
B 75.00
B
[ 2.953 ]
85.00
[ 3.346 ]
(47.00 )
[1.850 ]
8X 6.00
[ 0.236 ]
(90.00 )
[3.543 ]
LEGEND, THIS SHEET ONLY
ZONE 7:
NO COMPONENT PLACEMENT, STIFFENING PLATE CONTACT AREA
8 7 6 5 4 3 2 1
173
174
Figure 9-8.
B B
Primary and Secondary Side 3D Height Restriction Zones
A A
SECONDARY SIDE
DEPARTMENT SIZE DRAWING NUMBER REV
R
8 7 6 5 4 3 2 1
Boxed Processor Specifications
Figure 9-11. 4-Pin Fan Cable Connector (For Active Heat Sink)
Figure 9-12. 4-Pin Base Baseboard Fan Header (For Active Heat Sink)
Figure 9-13 illustrates the Unified Retention System (URS) and the Unified Backplate
Assembly. The URS is designed to extend air-cooling capability through the use of
larger heat sinks with minimal airflow blockage and bypass. URS retention transfers
load to the baseboard via the Unified Backplate Assembly. The URS spring, captive in
the heatsink, provides the necessary compressive load for the thermal interface
material.
All components of the URS heat sink solution will be captive to the heat sink and will
only require a Phillips screwdriver to attach to the Unified Backplate Assembly. When
installing the URS the screws should be tightened until they will no longer turn easily.
This should represent approximately 8 inch-pounds of torque. More than that may
damage the retention mechanism components.
Note: Actual boxed thermal solution may differ from this image, but installation is similar.
solution and does not support variable voltage control or 3-pin PWM control. See
Figure 9-14 and Table 9-1 through Table 9-3 for details on the 4-pin active heat sink
solution connectors.
The fan power header on the baseboard must be positioned to allow the fan heat sink
power cable to reach it. The fan power header identification and location must be
documented in the suppliers platform documentation, or on the baseboard itself. The
baseboard fan power header should be positioned within 177.8 mm [7 in.] from the
center of the processor socket.
Table 9-1. PWM Fan Frequency Specifications For 4-Pin Active Thermal Solution
Description Min Frequency Nominal Frequency Max Frequency Unit
PWM Control
21,000 25,000 28,000 Hz
Frequency Range
Figure 9-14. Fan Cable Connector Pin Out For 4-Pin Active Thermal Solution
Table 9-3. Fan Cable Connector Pin Out for 4-Pin Active Thermal Solution
Pin Number Signal Color
1 Ground Black
It is recommended that the ambient air temperature outside of the chassis be kept at
or below 35°C. The air passing directly over the processor thermal solution should not
be preheated by other system components. Meeting the processor’s temperature
specification is the responsibility of the system integrator.
This thermal solution is for use with 95 W and 130 W TDP processor SKUs.
Passive Configuration:
It is recommended that the ambient air temperature outside of the chassis be kept at
or below 35°C. The air passing directly over the processor thermal solution should not
be preheated by other system components. Meeting the processor’s temperature
specification is the responsibility of the system integrator.
This thermal solution is for use with processor SKUs no higher than 80 W.
Processors with a TDP of 95 W must provide a minimum airflow of 16 CFM at 0.40 in.
H2O (27.2 m3/hr at 99.5 Pa) of flow impedance. It is assumed that a TLA of 49°C is
met for 95 W processor installations. This requires a chassis design to limit the TRISE
at or below 14°C with an external ambient temperature of 35°C. Under these
conditions, only Thermal Profile B will be supported. If Thermal Profile A support is
desired for processors with a TDP of 95 W a 2U configuration with a chassis duct is
recommended. A TLA of <40°C is required. This requires a superior chassis design to
limit the TRISE below 5°C with an external ambient temperature of 35°C.
Processors with a TDP of 80 W or lower must provide a minimum airflow of 9.7 CFM at
0.20 in. H2O (16.5 m3/hr at 49.8 Pa) of flow impedance. It is assumed that a TLA of
49°C is met for these processor installations. This requires a chassis design to limit the
TRISE at or below 14°C with an external ambient temperature of 35°C.
Note: lease refer to the Intel® Xeon® Processor 5500/5600 Series Thermal / Mechanical
Design Guide for detailed mechanical drawings of the STS100P.
Boxed Processor
• Intel Xeon processor 5600 series processor
• Installation and warranty manual
• Intel Inside Logo