For Platforms Based on Mobile Intel® 4 Series Express Chipset Family
October 2009
Document Number: 322875-001EN
INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL® PRODUCTS. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR
OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. EXCEPT AS PROVIDED IN INTEL'S TERMS AND CONDITIONS
OF SALE FOR SUCH PRODUCTS, INTEL ASSUMES NO LIABILITY WHATSOEVER, AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING
TO SALE AND/OR USE OF INTEL PRODUCTS INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A PARTICULAR PURPOSE,
MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT, COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT.
UNLESS OTHERWISE AGREED IN WRITING BY INTEL, THE INTEL PRODUCTS ARE NOT DESIGNED NOR INTENDED FOR ANY APPLICATION IN WHICH T HE
FAILURE OF THE INTEL PRODUCT COULD CREATE A SITUATION WHERE PERSONAL INJURY OR DEATH MAY OCCUR.
Intel may make changes to specifications and product descriptions at any time, without notice. Designers must not rely on the absence or characteristics
of any features or instructions marked “reserved” or “undefined.” Intel reserves these for future definition and shall have no responsibility whatsoever for
conflicts or incompatibilities arising from future changes to them. The information here is subject to change without notice. Do not finalize a design with
this information.
The products described in this document may contain design defects or errors known as errata which may cause the product to deviate from published
specifications. Current characterized errata are available on request.
Contact your local Intel sales office or your distributor to obtain the latest specifications and before placing your product order.
64-bit computing on Intel architecture requires a computer system with a processor, chipset, BIOS, operating system, device drivers and applications
enabled for IntelÆ 64 architecture. Performance will vary depending on your hardware and software configurations. Consult with your system vendor for
more information.
Enhanced Intel SpeedStep® Technology for specified units of this processor are available. See the Processor Spec Finder at http://
processorfinder.intel.com or contact your Intel representative for more information.
Enabling Execute Disable Bit functionality requires a PC with a processor with Execute Disable Bit capability and a supporting operating system. Check
with your PC manufacturer on whether your system delivers Execute Disable Bit functionality.
Intel® Virtualization Technology requires a computer system with an enabled Intel® processor, BIOS, virtual machine monitor (VMM) and, for some
uses, certain platform software enabled for it. Functionality, performance or other benefits will vary depending on hardware and software configurations
and may require a BIOS update. Software applications may not be compatible with all operating systems. Please check with your application vendor.
• Processor offered in Micro-FCPGA packaging technology
• Execute Disable Bit support for enhanced security
• Half ratio support (N/2) for core to bus ratio
Datasheet
7
1.1Terminology
TermDefinition
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 level. For example, when
RESET# is low, a reset has been requested. Conversely, when NMI is high,
#
Front Side Bus
(FSB)
AGTL+
Storage
Conditions
Processor Core
Execute Disable
Bit
Intel® 64
Technology
a nonmaskable interrupt has occurred. In the case of signals where the
name does not imply an active state but describes part of a binary
sequence (such as address or data), the “#” symbol implies that the signal
is inverted. For example, D[3:0] = “HLHL” refers to a hex “A” , and D[3:0]#
= “LHLH” also refers to a hex “A” (H= High logic level, L= Low logic level).
Refers to the interface between the processor and system core logic (also
known as the chipset components).
Advanced Gunning Transceiver Logic. Used to refer to Assisted GTL+
signaling technology on some Intel processors.
Refers to a non-operational state. The processor may be installed in a
platform, in a tray, or loose. Processors may be sealed in packaging or
exposed to free air. Under these conditions, processor landings should not
be connected to any supply voltages, have any I/Os biased or receive any
clocks. Upon exposure to “free air” (i.e., unsealed packaging or a device
removed from packaging material) the processor must be handled in
accordance with moisture sensitivity labeling (MSL) as indicated on the
packaging material.
Processor core die with integrated L1 and L2 cache. All AC timing and signal
integrity specifications are at the pads of the processor core.
The Execute Disable bit allows memory to be marked as executable or nonexecutable, when combined with a supporting operating system. If code
attempts to run in non-executable memory the processor raises an error to
the operating system. This feature can prevent some classes of viruses or
worms that exploit buffer overrun vulnerabilities and can thus help improve
the overall security of the system. See the Intel® 64 and IA-32 Architectures Software Developer's Manuals for more detailed information.
64-bit memory extensions to the IA-32 architecture.
Introduction
Half ratio support
(N/2) for Core to
Bus ratio
TDPThermal Design Power.
V
CC
V
SS
8Datasheet
Intel Core 2 Duo processors and Intel Core 2 Extreme processors support
the N/2 feature that allows having fractional core-to-bus ratios. This featur e
provides the flexibility of having more frequency options and being able to
have products with smaller frequency steps.
The processor core power supply.
The processor ground.
Introduction
1.2References
Material and concepts available in the following documents may be beneficial when
reading this document.
Document
Intel® Pentium® Processor on 45-nm Technology Specification Update
Mobile Intel® 4 Series Express Chipset Family Datasheet320122
Mobile Intel® 4 Series Express Chipset Family Specification Update320123
Intel® I/O Controller Hub 9 (ICH9)/ I/O Controller Hub 9M (ICH9M)
Specification Update
Intel® 64 and IA-32 Architectures Software Developer's Manuals
Volume 1: Basic Architecture 253665
Volume 2A: Instruction Set Reference, A-M 253666
Volume 2B: Instruction Set Reference, N-Z 253667
Volume 3A: System Programming Guide 253668
Volume 3B: System Programming Guide 253669
NOTE: Contact your Intel representative for the latest revision of this document.
316972
316973
§
Document
Number
Datasheet
9
Introduction
10Datasheet
Low Power Features
2Low Power Features
2.1Clock Control and Low-Power States
The processor supports low-power states both at the individual core level and the
package level for optimal power management.
A core may independently enter the C1/AutoHALT, C1/MWAIT, C2, C3, low-power
states. When both cores coincide in a common core low-power state, the central power
management logic ensures the entire processor enters the respective package lowpower state by initiating a P_LVLx (P_LVL2, P_LVL3) I/O read to the GMCH.
The processor implements two software interfaces for requesting low-power states:
MWAIT instruction extensions with sub-state hints and P_L VLx reads to the ACPI P_BLK
register block mapped in the processor’s I/O address space. The P_LVLx I/O reads are
converted to equivalent MWAIT C-state requests inside the processor and do not
directly result in I/O reads on the processor FSB. The P_LVLx I/O Monitor address does
not need to be set up before using the P_LVLx I/O read interface. The sub-state hints
used for each P_LVLx read can be configured through the IA32_MISC_ENABLES model
specific register (MSR).
If a core encounters a GMCH break event while STPCLK# is asserted, it asserts the
PBE# output signal. Assertion of PBE# when STPCLK# is asserted indicates to system
logic that individual cores should return to the C0 state and the processor should return
to the Normal state.
Figure 1 shows the core low-power states and Figure 2 shows the package low-power
states for the processor. Table 1 maps the core low-power states to package low-power
states.
Datasheet
11
Figure 1.Core Low-Power States
C2
†
C0
Stop
Grant
Core state
break
P_LVL2 or
MWAIT(C2)
C3
†
Core
state
break
P_LVL3 or
MWAIT(C3)
C1/
MWAIT
Core state
break
MWAIT(C1)
C1/Auto
Halt
Halt break
HLT instruction
STPCLK#
de-asserted
STPCLK#
asserted
STPCLK#
de-asserted
STPCLK#
asserted
STPCLK#
de-asserted
STPCLK#
asserted
halt break = A20M# transition, INIT#, INTR, NMI, PREQ#, RESET#, SMI#, or APIC interrupt
core state break = (halt break OR Monitor event) AND STPCLK# high (not asserted)
† — STPCLK# assertion and de-assertion have no effect if a core is in C2 or C3.
Low Power Features
12Datasheet
Low Power Features
Stop
Grant
Snoop
Normal
Stop
Grant
STPCLK# asserted
Snoop
serviced
Snoop
occurs
Sleep
SLP# asserted
SLP# de-assertedSTPCLK# de-asserted
Deep
Sleep
DPSLP# asserted
DPSLP# de-asserted
Figure 2.Package Low-Power States
Table 1.Coordination of Core Low-Power States at the Package Level
This is the normal operating state for cores in the processor.
2.1.1.2Core C1/AutoHALT Powerdown State
C1/AutoHALT is a low-power state entered when a core executes the HALT instruction.
The processor core will transition to the C0 state upon occurrence of SMI#, INIT#,
LINT[1:0] (NMI, INTR), or FSB interrupt messages. RESET# will cause the processor to
immediately initialize itself.
A System Management Interrupt (SMI) handler will return execution to either Normal
state or the AutoHALT Powerdown state. See the Intel® 64 and IA-32 Architectures Software Developer's Manuals, Volume 3A/3B: System Programmer's Guide for more
information.
The system can generate a STPCLK# while the processor is in the AutoHALT
Powerdown state. When the system deasserts the STPCLK# interrupt, the processor
will return execution to the HALT state.
Datasheet
13
While in AutoHALT Powerdown state, the dual-core processor will process bus snoops
and snoops from the other core. The processor core will enter a snoopable sub-state
(not shown in Figure 1) to process the snoop and then return to the AutoHALT
Powerdow n state.
2.1.1.3Core C1/MWAIT Powerdown State
C1/MWAIT is a low-power state entered when the processor core executes the
MWAIT(C1) instruction. Processor behavior in the MWAIT state is identical to the
AutoHALT state except that Monitor events can cause the processor core to return to
the C0 state. See the Intel® 64 and IA-32 Architectures Software Developer's Manuals,
Volume 2A: Instruction Set Reference, A-M and Volume 2B: Instruction Set Reference,
N-Z, for more information.
2.1.1.4Core C2 State
Individual cores of the dual-core processor can enter the C2 state by initiating a P_LVL2
I/O read to the P_BLK or an MWAIT(C2) instruction, but the processor will not issue a
Stop-Grant Acknowledge special bus cycle unless the STPCLK# pin is also asserted.
While in the C2 state, the dual-core processor will process bus snoops and snoops from
the other core. The processor core will enter a snoopable sub-state (not shown in
Figure 1) to process the snoop and then return to the C2 state.
Low Power Features
2.1.1.5Core C3 State
Individual cores of the dual-core processor can enter the C3 state by initiating a P_LVL3
I/O read to the P_BLK or an MWAIT(C3) instruction. Before entering C3, the processor
core flushes the contents of its L1 caches into the processor’s L2 cache. Except for the
caches, the processor core maintains all its architectural states in the C3 state. The
Monitor remains armed if it is configured. 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 FSB or when the other core of the dual-core
processor accesses cacheable memory. The processor core will transition to the C0
state upon occurrence of a Monitor event, SMI#, INIT#, LINT[1:0] (NMI, INTR), or FSB
interrupt message. RESET# will cause the processor core to immediately initialize itself.
2.1.2Package Low-power State Descriptions
2.1.2.1Normal State
This is the normal operating state for the processor. The processor remains in the
Normal state when at least one of its cores is in the C0, C1/AutoHALT, or C1/MWAIT
state.
2.1.2.2Stop-Grant State
When the STPCLK# pin is asserted, each core of the dual-core processor enters the
Stop-Grant state within 20 bus clocks after the response phase of the processor-issued
Stop-Grant Acknowledge special bus cycle. Processor cores that are already in the C2,
C3, or C4 state remain in their current low-power state. When the STPCLK# pin is
deasserted, each core returns to its previous core low-power state.
Since the AGTL+ signal pins receive power from the FSB, these pins should not be
driven (allowing the level to return to V
termination resistors in this state. In addition, all other input pins on the FSB should be
driven to the inactive state.
14Datasheet
) for minimum power drawn by the
CCP
Low Power Features
RESET# causes the processor to immediately initialize itself, but the processor will stay
in Stop-Grant state. When RESET# is asserted by the system, the STPCLK#, SLP#,
DPSLP#, and DPRSTP# pins must be deasserted prior to RESET# deassertion as per AC
Specification T45. When re-entering the Stop-Grant state from the Sleep state,
STPCLK# should be deasserted after the deassertion of SLP# as per AC Specification
T75.
While in Stop-Grant state, the processor will service snoops and latch interrupts
delivered on the FSB. The processor will latch SMI#, INIT# and LINT[1:0] interrupts
and will service only one of each upon return to the Normal state.
The PBE# signal may be driven when the processor is in Stop-Grant state. PBE# will be
asserted if there is any pending interrupt or Monitor event latched within the processor.
Pending interrupts that are blocked by the EFLAGS.IF bit being clear will still cause
assertion of PBE#. Assertion of PBE# indicates to system logic that the entire processor
should return to the Normal state.
A transition to the Stop-Grant Snoop state occurs when the processor detects a snoop
on the FSB (see Section 2.1.2.3). A transition to the Sleep state (see Section 2.1.2.4)
occurs with the assertion of the SLP# signal.
2.1.2.3Stop-Grant Snoop State
The processor responds to snoop or interrupt transactions on the FSB while in StopGrant state by entering the Stop-Grant Snoop state. The processor will stay in this
state until the snoop on the FSB has been serviced (whether by the processor or
another agent on the FSB) or the interrupt has been latched. The processor returns to
the Stop-Grant state once the snoop has been serviced or the interrupt has been
latched.
2.1.2.4Sleep State
The Sleep state is a low-power state in which the processor maintains its context,
maintains the phase-locked loop (PLL), and stops all internal clocks. The Sleep state is
entered through assertion of the SLP# signal while in the Stop-Grant state. The SLP#
pin should only be asserted when the processor is in the Stop-Grant state. SLP#
assertions while the processor is not in the Stop-Grant state is out of specification and
may result in unapproved operation.
In the Sleep state, the processor is incapable of responding to snoop transactions or
latching interrupt signals. No transitions or assertions of signals (with the exception of
SLP#, DPSLP# or RESET#) are allowed on the FSB while the processor is in Sleep
state. Snoop events that occur while in Sleep state or during a transition into or out of
Sleep state will cause unpredictable behavior. Any transition on an input signal before
the processor has returned to the Stop-Grant state will result in unpredictable behavior.
If RESET# is driven active while the processor is in the Sleep state, and held active as
specified in the RESET# pin specification, then the processor will reset itself, ignoring
the transition through the Stop-Grant state. If RESET# is driven active while the
processor is in the Sleep state, the SLP# and STPCLK# signals should be deasserted
immediately after RESET# is asserted to ensure the processor correctly executes the
Reset sequence.
While in the Sleep state, the processor is capable of entering an even lower power
state, the Deep Sleep state, by asserting the DPSLP# pin (See Section 2.1.2.5). While
the processor is in the Sleep state, the SLP# pin must be deasserted if another
asynchronous FSB event needs to occur.
Datasheet
15
2.1.2.5Deep Sleep State
The Deep Sleep state is entered throug h assertion of the DPSL P# pin while in the Sleep
state. BCLK may be stopped during the Deep Sleep state for additional platform-level
power savings. BCLK stop/restart timings on appropriate GMCH-based platforms with
the CK505 clock chip are as follows:
• Deep Sleep entry: the system clock chip may stop/tristate BCLK within 2 BCLKs
of DPSLP# assertion. It is permissible to leave BCLK running during Deep Sleep.
• Deep Sleep exit: the system clock chip must drive BCLK to differential DC levels
within 2-3 ns of DPSLP# deassertion and start toggling BCLK within 10 BCLK
periods.
To re-enter the Sleep state, the DPSLP# pin must be deasserted. BCLK can be restarted after DPSLP# deassertion as described above. A period of 15 microseconds (to
allow for PLL stabilization) must occur before the processor can be considered to be in
the Sleep state. Once in the Sleep state, the SLP# pin must be deasserted to re-enter
the Stop-Grant state.
While in Deep Sleep state, the processor is incapable of responding to snoop
transactions or latching interrupt signals. No transitions of signals are allowed on the
FSB while the processor is in Deep Sleep state. When the processor is in Deep Sleep
state, it will not respond to interrupts or snoop transactions. Any transition on an input
signal before the processor has returned to Stop-Grant state will result in unpredictable
behavior.
Low Power Features
16Datasheet
Low Power Features
2.2Enhanced Intel SpeedStep® Technology
The processor features Enhanced Intel SpeedStep Technology. Following are the key
features of Enhanced Intel SpeedStep Technology:
• Multiple voltage and frequency operati ng points provide optimal performance at the
lowest power.
• Voltage and frequency selection is software-controlled by writing to processor
MSRs:
— If the target frequency is higher than the current frequency, V
in steps by placing new values on the VID pins, and the PLL then locks to the
new frequency.
— If the target frequency is lower than the current frequency, the PLL locks to the
new frequency and the VCC is changed through the VID pin mechanism.
— Software transitions are accepted at any time. If a previous transition is in
progress, the new transition is deferred until the previous transition completes.
• The processor controls voltage ramp rates internally to ensure glitch-free
transitions.
• Low transition latency and large number of transitions possible per second:
— Processor core (including L2 cache) is unavailable for up to 10 s during the
frequency transition.
— The bus protocol (BNR# mechanism) is used to block snooping.
• Improved Intel® Thermal Monitor mode:
— When the on-die thermal sensor indicates that the die temperature is too high
the processor can automatically perform a transition to a lower frequency and
voltage specified in a software-programmable MSR.
— The processor waits for a fixed time period. If the die temperature is down to
acceptable levels, an up-transition to the previous frequency and voltage point
occurs.
— An interrupt is generated for the up and down Intel Thermal Monitor transitions
enabling better system-level thermal management.
• Enhanced thermal management features:
— Digital Thermal Sensor and Out of Specification detection.
— Intel Thermal Monitor 1 (TM1) in addition to Intel Thermal Monitor 2 (TM2) in
case of unsuccessful TM2 transition.
— Dual-core thermal management synchronization.
is ramped up
CC
Each core in the dual-core processor implements an independent MSR for controlling
Enhanced Intel SpeedStep Technology, but both cores must operate at the same
frequency and voltage. The processor has performance state coordination logic to
resolve frequency and voltage requests from the two cores into a single frequency and
voltage request for the package as a whole. If both cores request the same frequency
and voltage, then the processor will transition to the requested common frequency and
voltage. If the two cores have different frequency and voltage requests, then the
processor will take the highest of the two frequencies and voltages as the resolved
request and transition to that frequency and voltage.
The processor also supports Dynamic FSB Frequency Switching and Intel Dynamic
Acceleration Technology mode on select SKUs. The operating system can take
advantage of these features and request a lower operating point called SuperLFM (due
to Dynamic FSB Frequency Switching) and a higher operating point Intel Dynamic
Acceleration Technology mode.
Datasheet
17
Low Power Features
2.3Extended Low-Power States
Extended low-power states (CXE) optimize for power by forcibly reducing the
performance state of the processor when it enters a package low-power state. Instead
of directly transitioning into the package low-power state, the enhanced package lowpower state first reduces the performance state of the processor by performing an
Enhanced Intel SpeedStep Technology transition down to the lowest operating point.
Upon receiving a break event from the package low-power state, control will be
returned to software while an Enhanced Intel SpeedStep Technology transition up to
the initial operating point occurs. The advantage of this feature is that it significantly
reduces leakage while in the Stop-Grant state.
Note:Long-term reliability cannot be assured unless all the Extended Low Power States are
enabled.
The processor implements two software interfaces for requesting enhanced package
low-power states: MWAIT instruction extensions with sub-state hints and via BIOS by
configuring IA32_MISC_ENABLES MSR bits to automatically promote package lowpower states to enhanced package low-power states.
Caution:Extended Stop-Grant must be enabled via the BIOSfor the processor to
Caution:Enhanced Intel SpeedStep Technology transitions are multistep processes
remain within specification. As processor technology changes, enabling the
extended low power states becomes increasingly crucial when building computer
systems. Maintaining the proper BIOS configuration is key to reliable, long-term
system operation. Not complying to this guideline may affect the long-term reliability of
the processor.
that require clocked control. These transitions cannot occur when the processor is in
the Sleep or Deep Sleep package low-power states since processor clocks are not
active in these states. The transition to the lowest operating point or back to the
original software-requested point may not be instantaneous. Furthermore, upon very
frequent transitions between active and idle states, the transitions may lag behind the
idle state entry resulting in the processor either executing for a longer time at the
lowest operating point or running idle at a high operating point. Observations and
analyses show this behavior should not significantly impact total power savings or
performance score while providing power benefits in most other cases.
18Datasheet
Low Power Features
2.4FSB Low Power Enhancements
The processor incorporates FSB low power enhancements:
• Dynamic FSB Power Down
• BPRI# control for address and control input buffers
• Dynamic Bus Parking
• Dynamic On-Die Termination disabling
•Low V
• Dynamic FSB frequency switching
The processor incorporates the DPWR# signal that controls the data bus input buffers
on the processor. The DPWR# signal disables the buffers when not used and activates
them only when data bus activity occurs, resulting in significant power savings with no
performance impact. BPRI# control also allows the processor address and control input
buffers to be turned off when the BPRI# signal is inactive. Dynamic Bus Parking allows
a reciprocal power reduction in GMCH address and control input buffers when the
processor deasserts its BR0# pin. The On-Die T ermination on the processor FSB buffers
is disabled when the signals are driven low, resulting in additional power savings. The
low I/O termination voltage is on a dedicated voltage plane independent of the core
voltage, enabling low I/O switching power at all times.
(I/O termination voltage)
CCP
2.5Processor Power Status Indicator (PSI-2) Signal
The processor incorporates the PSI# signal that is asserted when the processor is in a
reduced power consumption state. PSI# can be used to improve intermediate and light
load efficiency of the voltage regulator, resulting in platform power savings and
extended battery life. The algorithm that the processor uses for determining when to
assert PSI# is different from the algorithm used in previous mobile processors. PSI-2
functionality is expanded further to support three processor states:
• Both cores are in idle state
• Only one core active state
• Both cores are in active state
PSI-2 functionality improves overall voltage regulator efficiency over a wide power
range based on the C-state and P-state of the two cores. The combined C-state and Pstate of both cores are used to dynamically predict processor power.
The real-time power prediction is compared against a set of predefined and configured
valu es of CHH and CHL. CHH is indicative of the active C-state of both the cores and
CHL is indicative that only one core is in active C-state and the other core is in low
power core state. PSI-2# output is asserted upon crossing these thresholds indicating
that the processor requires lower power. The voltage regulator will adapt its power
output accordingly . Additionally the v oltage regulator may switch to a single phase and/
or asynchronous mode when the processor is idle and fused leakage limit is less than or
equal to the BIOS threshold value.
§
Datasheet
19
Low Power Features
20Datasheet
Electrical Specifications
3Electrical Specifications
3.1Power and Ground Pins
For clean, on-chip power distribution, the processor will have a large number of VCC
(power) and V
planes while all V
power and ground planes is recommended to reduce I*R drop. The processor V
must be supplied the voltage determined by the VID (Voltage ID) pins.
(ground) inputs. All power pins must be connected to V
SS
pins must be connected to system ground planes. Use of multiple
SS
3.2Decoupling Guidelines
Due to its large number of transistors and high internal clock speeds, the processor is
capable of generating large average current swings between low and full power states.
This may cause voltages on power planes to sag below their minimum values if bulk
decoupling is not adequate. Larger bulk storage, such as electrolytic capacitors, supply
current during longer lasting changes in current demand by the component, such as
coming out of an idle condition. Similarly, they act as a storage well for current when
entering an idle condition from a running condition. Care must be taken in the board
design to ensure that the voltage provided to the processor remains within the
specifications listed in the tables in Section 3.10. Failure to do so can result in timing
violations or reduced lifetime of the component.
3.2.1V
Decoupling
CC
V
regulator solutions need to provide bulk capacitance with a low Effective Series
CC
Resistance (ESR) and keep a low interconnect resistance from the regulator to the
socket. Bulk decoupling for the large current swings when the part is powering on, or
entering/exiting low-power states, should be provided by the voltage regulator solution
depending on the specific system design.
CC
power
CC
pins
3.2.2FSB AGTL+ Decoupling
The processors integrate signal termination on the die as well as incorporate high
frequency decoupling capacitance on the processor package. Decoupling must also be
provided by the system motherboard for proper AGTL+ bus operation.
3.2.3FSB Clock (BCLK[1:0]) and Processor Clocking
BCLK[1:0] directly controls the FSB interface speed as well as the core frequency of the
processor. As in previous-generation processors, the processor core frequency is a
multiple of the BCLK[1:0] frequency. The processor bus r atio multiplier will be set at its
default ratio at manufacturing. The processor uses a differential clocking
implementation.
Datasheet
21
Electrical Specifications
3.3Voltage Identification and Power Sequencing
The processor uses seven voltage identification pins,VID[6:0], to support automatic
selection of power supply voltages. The VID pins for the processor are CMOS outputs
driven by the processor VID circuitry. Table 2 specifies the voltage level corresponding
to the state of VID[6:0]. A 1 in the table refers to a high-voltage level and a 0 refers to
a low-voltage level.
Table 2.Voltage Identification Definition (Sheet 1 of 3)
The processor supports the THERMTRIP# signal for catastrophic thermal protection. An
external thermal sensor should also be used to protect the processor and the system
against excessive temperatures. Even with the activation of THERMTRIP#, which halts
all processor internal clocks and activity, leakage current can be high enough that the
processor cannot be protected in all conditions without the removal of power to the
processor. If the external thermal sensor detects a catastrophic processor temperature
of approximately 125°C (maximum), or if the THERMTRIP# signal is asserted, the V
supply to the processor must be turned off within 500 ms to prevent permanent silicon
damage due to thermal runaway of the processor. THERMTRIP# functionality is not
ensured if the PWRGOOD signal is not asserted, and during Deep Power Down
Technology State (C6).
3.5Reserved and Unused Pins
All RESERVED (RSVD) pins must remain unconnected. Connection of these pins to VCC,
V
, or to any other signal (including each other) can result in component malfunction
SS
or incompatibility with future processors. See Section 4.2 for a pin listing of the
processor and the location of all RSVD pins.
For reliable operation, always connect unused inputs or bidirectional signals to an
appropriate signal level. Unused active low AGTL+ inputs may be left as no-connects if
AGTL+ termination is provided on the processor silicon. Unused active high inputs
should be connected through a resistor to ground (V
unconnected. The TEST1,TEST2,TEST3,TEST4,TEST5,TEST6,TEST7 pins are used for
test purposes internally and can be left as “No Connects”.
SS
). Unused outputs can be left
CC
3.6FSB Frequency Select Signals (BSEL[2:0])
The BSEL[2:0] signals are used to select the frequency of the processor input clock
(BCLK[1:0]). These signals should be connected to the clock chip and the appropriate
chipset on the platform. The BSEL encoding for BCLK[1:0] is shown in Table 3.
The FSB signals have been combined into groups by buffer type in the following
sections. In this document, the term “AGTL+ Input” refers to the AGTL+ input group as
well as the AGTL+ I/O group when receiving. Similarly, “AGTL+ Output” refers to the
AGTL+ output group as well as the AGTL+ I/O group when driving.
With the implementation of a source-synchronous data bus, two sets of timing
parameters are specified. One set is for common clock signals, which are dependent
upon the rising edge of BCLK0 (ADS#, HIT#, HITM#, etc.), and the second set is for
the source-synchronous signals which are relative to their respective strobe lines (data
and address) as well as the rising edge of BCLK0. Asychronous signals are still present
(A20M#, IGNNE#, etc.) and can become active at any time during the clock cycle.
Table 4 identifies which signals are common clock, source synchronous, and
asynchronous.
Table 4.FSB Pin Groups
Signal GroupTypeSignals
Electrical Specifications
1
AGTL+ Common
Clock Input
AGTL+ Common
Clock I/O
AGTL+ Source
Synchronous
I/O
AGTL+ Strobes
CMOS InputAsynchronous
Open Drain
Output
Open Drain I/OAsynchronousPROCHOT#
CMOS OutputAsynchronousPSI#, VID[6:0], BSEL[2:0]
CMOS InputSynchronous to TCK TCK, TDI, TMS, TRST#
Open Drain
1.Refer to Chapter 4 for signal descriptions and termination requirements.
2.In processor systems where there is no debug port implemented on the system board,
these signals are used to support a debug port interposer. In systems with the debug port
implemented on the system board, these signals are no connects.
3.BPM[2:1]# and PRDY# are AGTL+ output-only signals.
4.PROCHOT# signal type is open drain output and CMOS input.
5.On-die termination differs from other AGTL+ signals.
3.8CMOS Signals
CMOS input signals are shown in Table 4. Legacy output FERR#, IERR# and other nonAGTL+ signals (THERMTRIP# and PROCHOT#) use Open Drain output buffers. These
signals do not have setup or hold time specifications in relation to BCLK[1:0]. However,
all of the CMOS signals are required to be asserted for more than four BCLKs for the
processor to recognize them. See Section 3.10 for the DC specifications for the CMOS
signal groups.
3.9Maximum Ratings
Table 5 specifies absolute maximum and minimum ratings only, which lie outside the
functional limits of the processor. Only within specified operation limits, can
functionality and long-term reliability be expected.
At conditions outside functional operation condition limits, but within absolute
maximum and minimum ratings, neither functionality nor long-term reliability can be
expected. If a device is returned to conditions within functional operation limits after
having been subjected to conditions outside these limits, but within the absolute
maximum and minimum ratings, the device may be functional, but with its lifetime
degraded depending on exposure to conditions exceeding the functional operation
condition limits.
At conditions exceeding absolute maximum and minimum ratings, neither functionality
nor long-term reliability can be expected. Moreover, if a device is subjected to these
conditions for any length of time then, when returned to conditions within the
functional operating condition limits, it will either not function, or its reliability will be
severely degraded.
Caution: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.
Table 5.Processor Absolute Maximum Ratings
SymbolParameterMinMax UnitNotes
T
STORAGE
T
STORAGE
V
CC
V
inAGTL+
V
inAsynch_CMOS
NOTES:
1.For functional operation, all processor electrical, signal quality, mechanical and thermal
specifications must be satisfied.
Processor Storage Temperature -40 85°C3,4,5
Processor Storage Temperature-25°C6
Any Processor Supply Voltage with
Respect to V
AGTL+ Buffer DC Input Voltage with
Respect to V
CMOS Buffer DC Input Voltage with
Respect to V
SS
SS
SS
-0.31.45V
-0.11.45V
-0.11.45V
1,2
Datasheet
27
2.Excessive overshoot or undershoot on any signal will likely re sult i n permanent damage to
the processor.
3.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 specifications.
4.This rating applies to the processor and does not include any tray or packaging.
5.Failure to adhere to this specif ication can affect the long-term reliability of the processor.
6.For Intel® Pentium® processors in 22x22 mm package.
3.10Processor DC Specifications
The processor DC specifications in this section are defined at the processor
core (pads) unless noted otherwise.
The tables list the DC specifications for the processor and are valid only while meeting
specifications for junction temperature, clock frequency, and input voltages. The
Highest Frequency Mode (HFM) and Lowest Frequency Mode (LFM) refer to the highest
and lowest core operating frequencies supported on the processor. Active mode load
line specifications apply in all states except in the Deep Sleep and Deeper Sleep states.
V
CC,BOOT
set the VID values. Unless specified otherwise, all specifications for the processor are at
TJ = 105 C. Read all notes associated with each parameter.
is the default voltage driven by the voltage regulator at power up in order to
Electrical Specifications
28Datasheet
Electrical Specifications
Table 6.Voltage and Current Specifications for the Pentium Processors
SymbolParameterMinTypMaxUnitNotes
V
CCHFM
V
CCLFM
V
CC,BOOT
V
CCP
V
CCA
I
CCDES
I
CC
I
AH,
I
SGNT
I
SLP
I
DSLP
dI
CC/DT
I
CCA
I
CCP
VCC at Highest Frequency Mode (HFM)0.91.2V1, 2
VCC at Lowest Frequency Mode (LFM)0.85—1.15V1, 2
Default V
Voltage for Initial Power Up—1.2—V2, 6
CC
AGTL+ Termination Voltage1.01.051.1V
PLL Supply Voltage1.4251.51.575V
ICC for Processors Recommended Design Target——47A10
ICC for Processors———
Processor
Number
T4500
T4400
T4300
T4200
Core Frequency/Voltage———
2.3 GHz & V
2.2 GHz & V
2.1 GHz & V
2.0 GHz & V
1.2 GHz & V
CCHFM
CCHFM
CCHFM
CCHFM
CCLFM
——
47
47
47
47
31.7
A3, 4
ICC Auto-Halt & Stop-Grant
HFM
LFM
——25.4
19.4
A3, 4
ICC Sleep
HFM
LFM
——24.7
19.2
A3, 4
ICC Deep Sleep
HFM
LFM
V
Power Supply Current Slew Rate at Processor
CC
Package Pin
ICC for V
I
for V
CCC
for V
I
CC
Supply——130mA
CCA
Supply before VCC Stable
CCP
Supply after VCC Stable
CCP
——22.9
A3, 4
18.5
——600mA/µs5, 7
——
4.5
2.5
A
A
8
9
NOTES:See next page.
Datasheet
29
Electrical Specifications
I
CC-CORE
max
{HFM|LFM}
V
CC-CORE
[V]
V
CC-CORE
nom {HFM|LFM}
+/-V
CC-CORE
Tolerance
= VR St. Pt. Error 1
/
V
CC-CORE, DC
min {HFM|LFM}
V
CC-CORE, DC
max {HFM|LFM}
V
CC-CORE
max {HFM|LFM}
V
CC-CORE
min {HFM|LFM}
10mV= RIPPLE
I
CC-CORE
[A]
0
Slope = -2.1 mV/A at package
VccSense, VssSense pins.
Differential Remote Sense required.
1.Each processor is programmed with a maximum valid voltage identification value (VID), which is set at
manufacturing and cannot be altered. Individual maximum VID values are calibrated during manufacturing
such that two processors at the same frequency may have different settings within the VID range. Note
that this differs from the VID employed by the processor during a power management event (Intel Thermal
Monitor 2, Enhanced Intel SpeedStep Technology, or Enhanced Halt State).
2.The voltage specifications are assumed to be measured across V
a 100-MHz bandwidth oscilloscope, 1.5-pF maximum probe capacitance, and 1-M minimum impedance.
CC_SENSE
and V
SS_SENSE
pins at socket with
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.
3.Specified at 105 °C T
4.Specified at the nominal V
5.Measured at the bulk capacitors on the motherboard.
6.V
tolerance shown in Figure 7 and Figure 8.
CC,BOOT
.
J
.
CC
7.Based on simulations and averaged over the duration of any change in current. Specified by design/
characterization at nominal V
8.This is a power-up peak current specification that is applicable when V
9.This is a steady-state I
10.Instantaneous current I
current will be less than maximum specified I
CC
CC_CORE_INST
. Not 100% tested.
CC
is high and V
CCP
current specification that is applicable when both V
of 57 A has to be sustained for short time (t
. VR OCP threshold should be high enough to support
CCDES
CCP
and V
CC_CORE
CC_CORE
) of 35 µs. Average
INST
is low.
are high.
current levels described herein.
Figure 3.Active VCC and ICC Loadline for Pentium Processors
1.Unless othe rwise noted, all specifications in th is table apply to all processor frequencie s.
2.Measured at 0.2 V.
3.V
4.For Vin between 0 V and V
is determined by value of the external pull-up resistor to V
OH
OH
.
CCP
.
5.Cpad includes die capacitance only. No package parasitics are included.
§
1
32Datasheet
Package Mechanical Specifications and Pin Information
4Package Mechanical
Specifications and Pin
Information
4.1Package Mechanical Specifications
The processor is available in 478-pin Micro-FCPGA packages. The package mechanical
dimensions are shown in Figure 9 through Figure 13.
The mechanical package pressure specifications are in a direction normal to the surface
of the processor. This protects the processor die from fracture risk due to uneven die
pressure distribution under tilt, stack-up tolerances and oth er similar conditions. These
specifications assume that a mechanical attach is designed specifically to load one type
of processor.
Moreover, the processor package substrate should not be used as a mechanical
reference or load-bearing surface for the thermal or mechanical solution.
Datasheet33
Package Mechanical Specifications and Pin Information
Package Mechanical Specifications and Pin Information
SignalsAssociated Strobe
REQ[4:0]#, A[16:3]# ADSTB[0]#
A[35:17]#ADSTB[1]#
4.3Alphabetical Signals Reference
Table 12.Signal Description (Sheet 1 of 8)
NameTypeDescription
A[35:3]# (Address) define a 2
space. In sub-phase 1 of the address phase, these pins transmit the
address of a transaction. In sub-phase 2, these pins tr an sm it
A[35:3]#
A20M#Input
ADS#
ADSTB[1:0]#
Input/
Output
Input/
Output
Input/
Output
transaction type information. These signals must connect the
appropriate pins of both agents on the processor FSB. A[35:3]# are
source-synchronous signals and are latched into the receiving
buffers by ADSTB[1:0]#. Address signals are used as straps, which
are sampled before RESET# is deasserted.
If A20M# (Address-20 Mask) is asserted, the processor masks
physical address bit 20 (A20#) before looking up a line in any
internal cache and before driving a read/write transaction on the
bus. Asserting A20M# emulates the 8086 processor's address
wrap-around at the 1-MB boundary. Assertion of A20M# is only
supported in real mode.
A20M# is an asynchronous signal. However, to ensure recognition
of this signal following an input/output write instruction, it must be
valid along with the TRDY# assertion of the corresponding input/
output Write bus transaction.
ADS# (Address Strobe) is asserted to indicate the validity of the
transaction address on the A[35:3]# and REQ[4:0]# pins. All bus
agents observe the ADS# activation to begin parity checking,
protocol checking, address decode, internal snoop, or deferred
reply ID match operations associated with the new transaction.
Address strobes are used to latch A[35:3]# and REQ[4:0]# on their
rising and falling edges. Strobes are associated with signals as
shown below.
36
-byte physical memory address
BCLK[1:0]Input
BNR#
BPM[2:1]#
BPM[3,0]#
Datasheet59
Input/
Output
Output
Input/
Output
The differential pair BCLK (Bus Clock) determines the FSB
frequency. All FSB agents must receive these signals to drive their
outputs and latch their inputs.
All external timing parameters are specified with respect to the
rising edge of BCLK0 crossing V
BNR# (Block Next Request) is used to assert a bus stall by any bus
agent who is unable to accept new bus transactions. During a bus
stall, the current bus owner cannot issue any new transactions.
BPM[3:0]# (Breakpoint Monitor) are breakpoint and performance
monitor signals. They are outputs from the processor that indicate
the status of breakpoints and programmable counters used for
monitoring processor performance. BPM[3:0]# should connect the
appropriate pins of all processor FSB agents.This includes debug or
performance monitoring tools.
CROSS
.
Package Mechanical Specifications and Pin Information
Quad-Pumped Signal Groups
Data Group
DSTBN#/
DSTBP#
DINV#
D[15:0]#00
D[31:16]#11
D[47:32]#22
D[63:48]#33
Table 12.Signal Description (Sheet 2 of 8)
NameTypeDescription
BPRI# (Bus Priority Request) is used to arbitrate for ownership of
the FSB. It must connect the appropriate pins of both FSB agents.
Observing BPRI# active (as asserted by the priority agent) causes
BPRI#Input
BR0#
BSEL[2:0]Output
COMP[3:0]Analog
Input/
Output
the other agent to stop issuing new requests, unless such requests
are part of an ongoing locked operation. The priority agent keeps
BPRI# asserted until all of its requests are completed, then releases
the bus by deasserting BPRI#.
BR0# is used by the processor to request the bus. The arbitration is
done between the processor (Symmetric Agent) and GMCH (High
Priority Agent).
BSEL[2:0] (Bus Select) are used to select the processor input clock
frequency. Table 3 defines the possible combinations of the signals
and the frequency associated with each combination. The required
frequency is determined by the processor, chipset and clock
synthesizer. All agents must operate at the same frequency.
COMP[3:0] must be terminated on the system board using
precision (1% tolerance) resistors.
D[63:0]# (Data) are the data signals. These signals provide a
64-bit data path between the FSB agents, and must connect the
appropriate pins on both agents. The data driver asserts DRDY# to
indicate a valid data transfer.
D[63:0]# are quad-pumped signals and will thus be driven four
times in a common clock period. D[63:0]# are latched off the
falling edge of both DSTBP[3:0]# and DSTBN[3:0]#. Each group of
16 data signals correspond to a pair of one DSTBP# and one
DSTBN#. The following table shows the grouping of data signals to
data strobes and DINV#.
D[63:0]#
DBR#Output
DBSY#
60Datasheet
Input/
Output
Input/
Output
Furthermore, the DINV# pins determine the polarity of the data
signals. Each group of 16 data signals corresponds to one DINV#
signal. When the DINV# signal is active, the corresponding data
group is inverted and therefore sampled active high.
DBR# (Data Bus Reset) is used only in processor 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. If a debug port is implemented in the system, DBR# is a no
connect in the system. DBR# is not a processor signal.
DBSY# (Data Bus Busy) is asserted by the agent responsible for
driving data on the FSB to indicate that the data bus is in use. The
data bus is released after DBSY# is deasserted. This signal must
connect the appropriate pins on both FSB agents.
Package Mechanical Specifications and Pin Information
DEFER# is asserted by an agent to indicate that a transaction
cannot be ensured in-order completion. Assertion of DEFER# is
DEFER#Input
normally the responsibility of the addressed memory or input/
output agent. This signal must connect the appropriate pins of both
FSB agents.
DINV[3:0]# (Data Bus Inversion) are source synchronous and
indicate the polarity of the D[63:0]# signals. The DINV[3:0]#
signals are activated when the data on the data bus is inverted. The
bus agent will invert the data bus signals if more than half the bits,
within the covered group, would change level in the next cycle.
DINV[3:0]#
DPRSTP#Input
DPSLP#Input
DPWR#
DRDY#
Input/
Output
Input/
Output
Input/
Output
DPRSTP#, when asserted on the platform, causes the processor to
transition from the Deep Sleep State to the Deeper Sleep state or
Deep Power Down Technology (C6) state. To return to the Deep
Sleep State, DPRSTP# must be deasserted. DPRSTP# is driven by
the ICH9M.
DPSLP# when asserted on the platform causes the processor to
transition from the Sleep State to the Deep Sleep state. T o return to
the Sleep State, DPSLP# must be deasserted. DPSLP# is driven by
the ICH9M.
DPWR# is a control signal used by the chipset to reduce power on
the processor data bus input buffers. The processor drives this pin
during dynamic FSB frequency switching
DRDY# (Data Ready) is asserted by the data driver on each data
transfer, indicating valid data on the data bus. In a multi-common
clock data transfer, DRDY# may be deasserted to insert idle clocks.
This signal must connect the appropriate pins of both FSB agents.
Data strobe used to latch in D[63:0]#.
.
DSTBN[3:0]#
Datasheet61
Input/
Output
Package Mechanical Specifications and Pin Information
FERR# (Floating-point Error)/PBE# (Pending Break Event) is a
multiplexed signal and its meaning is qualified with STPCLK#. When
STPCLK# is not asserted, FERR#/PBE# indicates a floating point
when the processor detects an unmasked floating-point error.
FERR# is similar to the ERROR# signal on the Intel® 387
coprocessor, and is included for compatibility with systems using
Microsoft MS-DOS*-type floating-point error reporting. When
STPCLK# is asserted, an assertion of FERR#/PBE# indicates that
the processor has a pending break event waiting for service. The
assertion of FERR#/PBE# indicates that the processor should be
returned to the Normal state. When FERR#/PBE# is asserted,
indicating a break event, it will remain asserted until STPCLK# is
deasserted. Assertion of PREQ# when STPCLK# is active will also
cause an FERR# break event.
For additional information on the pending break event functionality,
including identification of support of the feature and enable/disable
information, refer to Volumes 3A and 3B of the Intel® 64 and IA-32
Architectures Software Developer's Manuals and the Intel®
Processor Identification and CPUID Instruction application note.
GTLREF determines the signal reference level for AGTL+ input pins.
GTLREF should be set at 2/3 V
receivers to determine if a signal is a logical 0 or logical 1.
. GTLREF is used by the AGTL+
CCP
HIT# (Snoop Hit) and HITM# (Hit Modified) convey transaction
snoop operation results. Either FSB agent may assert both HIT#
and HITM# together to indicate that it requires a snoop stall that
can be continued by reasserting HIT# and HITM# together.
IERR# (Internal Error) is asserted by the processor as the result of
an internal error. Assertion of IERR# is usually accompanied by a
SHUTDOWN transaction on the FSB. This transaction may optionally
be converted to an external error signal (e.g., NMI) by system core
logic. The processor will keep IERR# asserted until the assertion of
RESET#, BINIT#, or INIT#.
IGNNE# (Ignore Numeric Error) is asserted to force the processor
to ignore a numeric error and continue to execute non-control
floating-point instructions. If IGNNE# is deasserted, the processor
generates an exc eption on a non-co ntrol floating-point instruction i f
a previous floating-point instruction caused an error. IGNNE# has
no effect when the NE bit in control register 0 (CR0) is set.
IGNNE# is an asynchronous signal. However, to ensure recognition
of this signal following an input/output write instruction, it must be
valid along with the TRDY# assertion of the corresponding input/
output Write bus transaction.
62Datasheet
Package Mechanical Specifications and Pin Information
Table 12.Signal Description (Sheet 5 of 8)
NameTypeDescription
INIT# (Initialization), when asserted, resets integer registers inside
the processor without affecting its internal caches or floating-point
registers. The processor then begins execution at the power-on
Reset vector configured during power-on configuration. The
processor continues to handle snoop requests during INIT#
INIT#Input
LINT[1:0]Input
LOCK#
PRDY#Output
PREQ#Input
PROCHOT#
PSI#Output
Input/
Output
Input/
Output
assertion. INIT# is an asynchronous signal. However, to ensure
recognition of this signal following an input/output write instruction,
it must be valid along with the TRDY# assertion of the
corresponding input/output write bus transaction. INIT# must
connect the appropriate pins of both FSB agents.
If INIT# is sampled active on the active-to-inactive transition of
RESET#, then the processor executes its Built-in Self-Test (BIST)
LINT[1:0] (Local APIC Interrupt) must connect the appropriate pins
of all APIC Bus agents. When the APIC is disabled, the LINT0 signal
becomes INTR, a maskable interrupt request signal, and LINT1
becomes NMI, a nonmaskable interrupt. INTR and NMI are
backward-compatible with the signals of those names on the
Pentium processor. Both signals are asynchronous.
Both of these signals must be software configured via BIOS
programming of the APIC register space to be used either as NMI/
INTR or LINT[1:0]. Because the APIC is enabled by default after
Reset, operation of these pins as LINT[1:0] is the default
configuration.
LOCK# indicates to the system that a transaction must occur
atomically. This signal must connect the appropriate pins of both
FSB agents. For a locked sequence of transactions, LOCK# is
asserted from the beginning of the first transaction to the end of
the last transaction.
When the priority agent asserts BPRI# to arbitrate for ownership of
the FSB, it will wait until it observes LOCK# deasserted. This
enables symmetric agents to retain ownership of the FSB
throughout the bus locked operation and ensure the atomicity of
lock.
Probe Ready signal used by debug tools to determine processor
debug readiness.
Probe Request signal used by debug tools to request debug
operation of the processor.
As an output, PROCHOT# (Processor Hot) 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 (TCC) has
been activated, if enabled. As an input, assertion of PROCHOT# by
the system will activate the TCC, if enabled. The TCC will remain
active until the system deasserts PROCHOT#.
By default PROCHOT# is configured as an output. The processor
must be enabled via the BIOS for PROCHOT# to be configured as
bidirectional.
This signal may require voltage translation on the motherboard.
Processor Power Status Indicator signal. This signal is asserted
when the processor is both in the normal stat e (HFM to LF M) and in
lower power states (Deep Sleep and Deeper Sleep).
Datasheet63
Package Mechanical Specifications and Pin Information
Table 12.Signal Description (Sheet 6 of 8)
NameTypeDescription
PWRGOOD (Power Good) is a processor input. The processor
requires this signal to be a clean indication that the clocks and
power supplies are stable and within their specifications. ‘Clean’
implies that the signal remains low (capable of sinking leakage
PWRGOODInput
REQ[4:0]#
Input/
Output
RESET#Input
RS[2:0]#Input
Reserved/
RSVD
No
Connect
SLP#Input
SMI#Input
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 .
The PWRGOOD 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.
REQ[4:0]# (Request Command) must connect the appropriate pins
of both FSB agents. They are asserted by the current bus owner to
define the currently active transaction type. These signals are
source synchronous to ADSTB[0]#.
Asserting the RESET# signal resets the processor to a known state
and invalidates its internal caches with ou t writ ing bac k any of their
contents. For a power-on Reset, RESET# must stay active for at
least two milliseconds after V
proper specifications. On observing active RESET#, both FSB
agents will deassert their outputs within two clocks. All processor
straps must be valid within the specified setup time before RESET#
is deasserted.
There is a 55 (nominal) on die pull-up resistor on this signal.
RS[2:0]# (Response St atus) are driv en b y the res ponse agent (t he
agent responsible for completion of the current transaction), and
must connect the appropriate pins of both FSB agents.
These pins are RESERVED and must be left unconnected on the
board. However, it is recommended that routing channels to these
pins on the board be kept open for possible future use.
SLP# (Sleep), when asserted in Stop-Grant state, causes the
processor to enter the Sleep state. During Sleep state, the
processor stops providing internal clock signals to all units, leaving
only the Phase-Locked Loop (PLL) still operating. Processors in this
state will not recognize snoops or interrupts. The processor will
recognize only assertion of the RESET# signal, deassertion of SLP#,
and removal of the BCLK input while in Sleep state. If SLP# is
deasserted, the processor exits Sleep state and returns to StopGrant state, restarting its internal clock signals to the bus and
processor core units. If DPSLP# is asserted while in the Sleep state,
the processor will exit the Sleep state and transition to the Deep
Sleep state.
SMI# (System Management Interrupt) is asserted asynchronously
by system logic. On accepting a System Management Interrupt, the
processor saves the current state and enters System Management
Mode (SMM). An SMI Acknowledge transaction is issued and the
processor begins program execution from the SMM handler.
If an SMI# is asserted during the deassertion of RESET#, then the
processor will tristate its outputs.
and BCLK have reached their
CC
64Datasheet
Package Mechanical Specifications and Pin Information
Table 12.Signal Description (Sheet 7 of 8)
NameTypeDescription
STPCLK# (Stop Clock), when asserted, causes the processor to
enter a low power Stop-Grant state. The processor issues a StopGrant Acknowledge transaction, and stops providing internal clock
signals to all processor core units except the FSB and APIC units.
VSSInputProcessor core ground node.
VCCAInputVCCA provides isolated power for the internal processor core PLLs
VCCPInputProcessor I/O Power Supply.
The processor continues to snoop bus transactions and service
interrupts while in Stop-Grant state. When STPCLK# is deasserted,
the processor restarts its internal clock to all units and resume s
execution. The assertion of STPCLK# has no effect on the bus
clock; STPCLK# is an asynchronous input.
TCK (Test Clock) provides the clock input for the processor T est Bus
(also known as the Test Access Port).
TDI (Test Data In) transfers serial test data into the processor. TDI
provides the serial input needed for JTAG specification support.
TDO (Test Data Out) transfers serial test data out of the processor.
TDO provides the serial output needed for JTAG specification
support.
Refer to the appropriate platform design guide for further TEST1,
TEST2, TEST3, TEST4, TEST5, TEST6 and TEST7 termination
requirements and implementation details.
The processor protects itself from catastrophic overheating by use
of an internal thermal sensor. This sensor is set well above the
normal operating temperature to ensure that there are no false
trips. The processor will stop all execution when the junction
temperature exceeds approximately 125 °C. This is signalled to the
system by the THERMTRIP# (Thermal Trip) pin.
TMS (Test Mode Select) is a JTAG specification support signal used
by debug tools.
TRDY# (Target Ready) is asserted by the target to indicate that it is
ready to receive a write or implicit writeback data transfer. TRDY#
must connect the appropriate pins of both FSB agents.
TRST# (Test Reset) resets the Test Access Port (TAP) logic. TRST#
must be driven low during power on Reset.
.
Datasheet65
Package Mechanical Specifications and Pin Information
Table 12.Signal Description (Sheet 8 of 8)
NameTypeDescription
VCCSENSE together with VSSSENSE are voltage feedback signals
VCCSENSEOutpu t
VID[6:0]Output
VSSSENSEOutput
that control the 2.1 m loadline at the processor die. It should be
used to sense voltage near the silicon with little noise.
VID[6:0] (Voltage ID) pi ns are used to support automatic selection
of power supply voltages (V
of processors, these are CMOS signals that are driven by the
processor. The voltage supply for these pins must be valid before
the VR can supply V
must be disabled until the voltage supply for the VID pins becomes
valid. The VID pins are needed to support the processor voltage
specification variations. See Table 2 for definitions of these pins.
The VR must supply the voltage that is requested by the pins, or
disable itself.
VSSSENSE
that control the 2.1-m loadline at the processor die. It should be
used to sense ground near the silicon with little noise.
). Unlike some previous generations
CC
to the processor. Conversely, the VR output
CC
together with VCCSENSEare voltage feedback signals
§
66Datasheet
Thermal Specifications and Design Considerations
5Thermal Specifications and
Design Considerations
A complete thermal solution includes both component and system-level thermal
management features. To allow for the optimal operation and long-term reliability of
Intel processor-based systems, the system/processor thermal solution should be
designed so the processor remains within the minimum and maximum junction
temperature (T
listed in the tables below
) specifications at the corresponding thermal design power (TDP) value
J
Caution:Operating the processor outside these operating limits may result in permanent
damage to the processor and potentially other components in the system.
Datasheet
67
Thermal Specifications and Design Considerations
Table 13.Power Specifications for the Pentium Processors
Symbol
TDP
Processor
Number
T4500
T4400
T4300
T4200
Core Frequency & Voltage
2.30 GHz & V
2.20 GHz & V
2.1 GHz & V
2.0 GHz & V
1.2 GHz & V
CCHFM
CCHFM
CCHFM
CCHFM
CCLFM
Thermal Design
Power
35
35
35
17
UnitNotes
W1, 4, 5
SymbolParameterMin Typ MaxUnitNotes
P
AH,
P
SGNT
Auto Halt, Stop Grant Power
at V
CCHFM
at V
CCLFM
——13.9
7.4
W2, 6
Sleep Power
P
SLP
at V
at V
CCHFM
CCLFM
——13.1
7.1
W2, 6
Deep Sleep Power
P
T
DSLP
J
at V
at V
CCHFM
CCLFM
——5.5
3.2
Junction Temperature0—105C3, 4
W2, 5, 7
NOTES:
1.The TDP specification should be used to design the processor thermal solution. The TDP is
not the maximum theoretical power the processor can generate.
2.Not 100% tested. These power specifications are determined by characterization of the
processor currents at higher temperatures and extrapolating the values for the
temperature indicated.
3.As measured by the activation of the on-die Intel Thermal Monitor. The Intel Thermal
Monitor’s automatic mode is used to indicate that the maximum T
4.The Intel Thermal Monitor automatic mode must be enabled for the processor to operate
within specifications.
5.At Tj of 105
6.At Tj of 50
7.At Tj of 35
o
C
o
C
o
C
has been reached.
J
68Datasheet
Thermal Specifications and Design Considerations
5.1Monitoring Die Temperature
The processor incorporates three methods of monitoring die temperature:
•Thermal Diode
•Intel® Thermal Monitor
• Digital Thermal Sensor
5.1.1Thermal Diode
Intel’s processors utilize an SMBus thermal sensor to read back the voltage/current
characteristics of a substrate PNP transistor. Since these characteristics are a function
of temperature, these parameters can be used to calculate silicon temperature values.
For older silicon process technologies, it is possible to simplify the voltage/current and
temperature relationships by treating the substrate transistor as though it were a
simple diffusion diode. In this case, the assumption is that the beta of the transistor
does not impact the calculated temperature values. The resultant “diode” model
essentially predicts a quasi linear relationship between the base/emitter voltage
differential of the PNP transistor and the applied temperature (one of the
proportionality constants in this relationship is processor specific, and is known as the
diode ideality factor). Realization of this relationship is accomplished with the SMBus
thermal sensor that is connected to the transistor.
The processor, howe ver, is built on Intel’s advanced 45-nm processor technology. Due
to this new processor technology, it is no longer possible to model the substrate
transistor as a simple diode. To accurately calculate silicon temperature use a full bipolar junction transistor-type model. In this model, the voltage/current and
temperature characteristics include an additional process dependant parameter which
is known as the transistor “beta”. System designers should be aware that the current
thermal sensors may not be configured to account for “beta” and should work with their
SMB thermal sensor vendors to ensure they have a part capable of reading the thermal
diode in BJT model.
Offset between the thermal diode-based temperature reading and the Intel Thermal
Monitor reading may be characterized using the Intel Thermal Monitor’s Automatic
mode activation of the thermal control circuit. This temperature offset must be
considered when using the processor thermal diode to implement power management
events. This offset is different than the diode Toffset value programmed into the
processor Model-Specific Register (MSR).
Table 14 and Table 15 provide the diode interface and transistor model specifications.
exemplified by the equation for the collector current:
I
= IS * (e
C
where I
base emitter junction (same nodes as VD), k = Boltzmann Constant, and T = absolute
temperature (Kelvin).
= saturation current, q = electronic charge, VBE = voltage across the transistor
S
qVBE/nQkT
–1)
5.1.2Intel® Thermal Monitor
The Intel Thermal Monitor helps control the processor temperature by activating the
TCC (Thermal Control Circuit) when the processor silicon reaches its maximum
operating temperature. The temperature at which the Intel Thermal Monitor activates
the TCC is not user configurable. Bus traffic is snooped in the normal manner and
interrupt requests are latched (and serviced during the time that the clocks are on)
while the TCC is active.
With a properly designed and characterized thermal solution, 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 minor and hence not detectable. 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 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.
The Intel Thermal Monitor controls the processor temperature by modulating (starting
and stopping) the processor core clocks or by initiating an Enhanced Intel SpeedStep
Technology transition when the processor silicon reaches its maximum operating
temperature. The Intel Thermal Monitor uses two modes to activate the TCC: automatic
mode and on-demand mode. If both modes are activated, automatic mode takes
precedence.
There are two automatic modes called Intel Thermal Monitor 1 (TM1) and Intel Thermal
Monitor 2 (TM2). These modes are selected by writing values to the MSRs of the
processor. After automatic mode is enabled, the TCC will activate only when the
internal die temperature reaches the maximum allowed value for operation.
When TM1 is enabled and a high temperature situation exists, the clocks will be
modulated by alternately turning the clocks off and on at a 50% duty cycle. Cy cle times
are processor speed-dependent and will decrease linearly as processor core frequencies
increase. Once the temperature has returned to a non-critical level, modulation ceases
and TCC goes inactive. A small amount of hysteresis has been included to prevent rapid
70Datasheet
Thermal Specifications and Design Considerations
active/inactive transitions of the TCC when the processor temperature is near the trip
point. The duty cycle is factory configured and cannot be modified. Also, automatic
mode does not require any additional hardware, software drivers, or interrupt handling
routines. Processor performance will be decreased by the same amount as the duty
cycle when the TCC is active.
When TM2 is enabled and a high temperature situation exists, the processor will
perform an Enhanced Intel SpeedStep Technology transition to the LFM. When the
processor temperature drops below the critical level, the processor will make an
Enhanced Intel SpeedStep Technology transition to the last requested operating point.
The processor also supports Enhanced Multi-Threaded Thermal Monitoring (EMTTM).
EMTTM is a processor feature that enhances TM2 with a processor throttling algorithm
known as Adaptive TM2. Adaptive TM2 transitions to intermediate operating points,
rather than directly to the LFM, once the processor has reached its thermal limit and
subsequently searches for the highest possible operating point. Please ensure this
feature is enabled and supported in the BIOS. Also with EMTTM enabled, the oper ating
system can request the processor to throttling to any point between Intel
Dynamic Acceleration Technology frequency and SuperLFM frequency as long
as these features are enabled in the BIOS and supported by the processor.
The Intel Thermal Monitor automatic mode and Enhanced Multi-Threaded
Thermal Monitoring must be enabled through BIOS for the processor to be
operating within specifications. Intel recommends TM1 and TM2 be enabled on the
processors.
TM1, TM2 and EMTTM features are collectively referred to as Adaptive Thermal
Monitoring features.
TM1 and TM2 can co-exist within the processor. If both TM1 and TM2 bits are enabled in
the auto-throttle MSR, TM2 takes precedence over TM1. However, if Force TM1 over
TM2 is enabled in MSRs via BIOS and TM2 is not sufficient to cool the processor below
the maximum operating temperature, then TM1 will also activate to help cool down the
processor.
If a processor load-based Enhanced Intel SpeedStep Technology transition (through
MSR write) is initiated when a TM2 period is active, there are two possible results:
1. If the processor load-based Enhanced Intel SpeedStep T echnology transition target
frequency is higher than the TM2 transition-based target frequency, the processor
load-based transition will be deferred until the TM2 event has been completed.
2. If the processor load-based Enhanced Intel SpeedStep T echnology transition target
frequency is lower than the TM2 transition-based target frequency, the processor
will transition to the processor load-based Enhanced Intel SpeedStep Technology
target frequency point.
The TCC may also be activated via on-demand mode. If bit 4 of the ACPI Intel Thermal
Monitor control register is written to a 1, the TCC will be activated immediately
independent of the processor temperature. When using on-demand mode to activate
the TCC, the duty cycle of the clock modulation is programmable via bits 3:1 of the
same ACPI Intel Thermal Monitor control register. In automatic mode, the duty cycle is
fixed at 50% on, 50% off, however in on-demand mode, the duty cycle can be
programmed from 12.5% on/ 87.5% off, to 87.5% on/12.5% off in 12.5% increments.
On-demand mode may be used at the same time automatic mode is enabled, however,
if the system tries to enable the TCC via on-demand mode at the same time autom atic
mode is enabled and a high temperature condition exists, automatic mode will take
precedence.
An external signal, PROCHOT# (processor hot) is asserted when the processor detects
that its temperature is above the thermal trip point. Bus snooping and interrupt
latching are also active while the TCC is active.
Datasheet
71
Thermal Specifications and Design Considerations
Besides the thermal sensor and thermal control circuit, the Intel Thermal Monitor also
includes one ACPI register, one performance counter register, three MSR, and one I/O
pin (PROCHOT#). All are available to monitor and control the state of the Intel Thermal
Monitor feature. The Intel Thermal Monitor can be configured to generate an interrupt
upon the assertion or deassertion of PROCHOT#.
PROCHOT# will not be asserted when the processor is in the Stop Grant,
Sleep, Deep Sleep, and Deeper Sleep low-power states, hence the thermal
diode reading must be used as a safeguard to maintain the processor junction
temperature within maximum specification. If the platform thermal solution is not
able to maintain the processor junction temperature within the maximum specification,
the system must initiate an orderly shutdown to prevent damage. If the processor
enters one of the above low-power states with PROCHOT# already asserted,
PROCHOT# will remain asserted until the processor exits the low-power state and the
processor junction temperature drops below the thermal trip point. However,
PROCHOT# will de-assert for the duration of Deep Power Down Technology state (C6)
residency.
If Thermal Monitor automatic mode is disabled, the processor will be operating out of
specification. Regardless of enabling the automatic or on-demand modes, in the event
of a catastrophic cooling failure, the processor will automatically shut down when the
silicon has reached a temperature of approximately 125 °C. At this point the
THERMTRIP# signal will go active. THERMTRIP# activation is independent of processor
activity and does not generate any bus cycles. When THERMTRIP# is asserted, the
processor core voltage must be shut down within the time specified in Chapter 3
.
In all cases the Intel Thermal Monitor feature must be enabled for the processor to
remain within specification.
5.1.3Digital Thermal Sensor
The processor also contains an on-die Digital Thermal Sensor (DTS) that can be read
via an MSR (no I/O interface). Each core of the processor will have a unique digital
thermal sensor whose temperature is accessible via the processor MSRs. The DTS is the
preferred method of reading the processor die temperature since it can be located
much closer to the hottest portions of the die and can thus more accurately track the
die temperature and potential activation of processor core clock modulation via the
Thermal Monitor. The DTS is only valid while the processor is in the normal operating
state (the Normal package level low-power state).
Unlike traditional thermal devices, the DTS outputs a temperature relative to the
maximum supported operating temperature of the processor (T
responsibility of software to convert the relative temperature to an absolute
temperature. The temperature returned by the DTS will always be at or below T
Catastrophic temperature conditions are detectable via an Out Of Specification status
bit. This bit is also part of the DTS MSR. When this bit is set, the processor is operating
out of specification and immediate shutdown of the system should occur. The processor
operation and code execution is not ensured once the activation of the Out of
Specification status bit is set.
The DTS-relative temperature readout corresponds to the Thermal Monitor (TM1/TM2)
trigger point. When the DTS indicates maximum processor core temperature has been
reached, the TM1 or TM2 hardware thermal control mechanism will activate. The DTS
and TM1/TM2 temperature may not correspond to the thermal diode reading since the
thermal diode is located in a separate portion of the die and thermal gradient between
the individual core DTS. Additionally, the thermal gradient from DTS to thermal diode
can vary substantially due to changes in processor power, mechanical and thermal
attach, and software application. The system designer is required to use the DTS to
ensure proper operation of the processor within its temperature operating
specifications.
J,max
). It is the
J,max
.
72Datasheet
Thermal Specifications and Design Considerations
Changes to the temperature can be detected via two programmable thresholds located
in the processor MSRs. These thresholds have the capability of generating interrupts
via the core's local APIC. Refer to the Intel® 64 and IA-32 Architectures Software Developer's Manuals for specific register and programming details.
5.2Out of Specification Detection
Overheat detection is performed by monitoring the processor temperature and
temperature gradient. This feature is intended for graceful shutdown before the
THERMTRIP# is activated. If the processor’s TM1 or TM2 are triggered and the
temperature remains high, an Out Of Spec status and sticky bit are latched in the
status MSR register and generates a thermal interrupt.
5.3PROCHOT# Signal Pin
An external signal, PROCHOT# (processor hot), is asserted when the processor die
temperature has reached its maximum operating temperature. If TM1 or TM2 is
enabled, then the TCC will be active when PROCHOT# is asserted. The processor can
be configured to generate an interrupt upon the assertion or deassertion of
PROCHOT#. Refer to the an interrupt upon the assertion or deassertion o f PROCHOT#.
Refer to the Intel® 64 and IA-32 Architectures Software Developer's Manuals for
specific register and programming details.
The processor implements a bi-directional PROCHOT# capability to allow system
designs to protect various components from overheating situations. The PROCHOT#
signal is bi-directional in that it can either signal when the processor 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.
Only a single PROCHOT# pin exists at a package level of the processor. When either
core's thermal sensor trips, PROCHOT# signal will be driven by the processor package.
If only TM1 is enabled, PROCHOT# will be asserted regardless of which core is above its
TCC temperature trip point, and both cores will have their core clocks modulated. If
TM2 is enabled then, regardless of which core(s) are above the TCC temperature trip
point, both cores will enter the lowest programmed TM2 performance state. It is
important to note that Intel recommends both TM1 and TM2 to be enabled.
When PROCHOT# is driven by an external agent, if only TM1 is enabled on both cores,
then both processor cores will have their core clocks modulated. If TM2 is enabled on
both cores, then both processor cores will enter the lowest programmed TM2
performance state. It should be noted that Force TM1 on TM2, enabled via BIOS, does
not have any effect on external PROCHOT#. If PROCHOT# is driven by an external
agent when TM1, TM2, and Force TM1 on TM2 are all enabled, then the processor will
still apply only TM2.
PROCHOT# may be used for thermal protection of voltage regulators (VR). System
designers can create a circuit to monitor the VR temperature and activate the TCC
when the temperature limit of the VR is reached. By asserting PROCHOT# (pulled-low)
and activating the TCC, the VR will cool down as a result of reduced processor power
consumption. Bi-directional PROCHOT# can allow VR thermal designs to target
maximum sustained current instead of maximum current. Systems should still provide
proper cooling for the VR and rely on bi-directional 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 TDP. With a properly designed and characterized thermal solution, it is
anticipated that bi-directional PROCHOT# would only be asserted for very short periods
Datasheet
73
Thermal Specifications and Design Considerations
of time when running the most power-intensive applications. An under-designed
thermal solution that is not able to prevent excessive assertion of PROCHOT# in the
anticipated ambient environment may cause a noticeable performance loss.
§
74Datasheet
Loading...
+ hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.