Intel technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service
activation. Learn more at Intel.com, or from the OEM or retailer.
No computer system can be absolutely secure. Intel does not assume any liability for lost or stolen data or systems or any
damages resulting from such losses.
You may not use or facilitate the use of this document in connection with any infringement or other legal analysis concerning Intel
products described herein. You agree to grant Intel a non-exclusive, royalty-free license to any patent claim thereafter drafted
which includes subject matter disclosed herein.
No license (express or implied, by estoppel or otherwise) to any intellectual property rights is granted by this document.
The products described 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.
Intel disclaims all express and implied warranties, including without limitation, the implied warranties of merchantability, fitness for
a particular purpose, and non-infringement, as well as any warranty arising from course of performance, course of dealing, or
usage in trade.
Copies of documents which have an order number and are referenced in this document may be obtained by calling 1-800-5484725 or by visiting www.intel.com/design/literature.htm.
Intel, Xeon, Pentium, Intel Core, and the Intel logo are trademarks of Intel Corporation in the U. S. and other countries.
*Other names and brands may be claimed as the property of others.
Intel® Xeon® Processor E5 v3 Product Family 3
Processor Specification Update
February 2016
Revision History
DateRevisionDescription
February 2016-010• Added errata HSE106 - HSE109
August 2015-009• Added errata HSE102 - HSE105
July 2015-008• Added errata HSE95 - HSE101
June 2015-007
March 2015-006• Added errata HSE88 - HSE92
January 2015-005
December 2014-004• Added errata HSE81 - HSE85
November 2014-003
October 2014-002• Updated the SKU list
September 2014-001• Initial Release
• Added errata HSE93 - HSE94
• Deleted erratum HSE79. Duplicate
• Added errata HSE86 - HSE87
•Added E5-EN/4S SKU info to Table 1, 2, 3.
• Added errata HSE77 - HSE80
• Added Document Change - Statement of Volatility
§
4Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
Preface
This document is an update to the specifications contained in the Affected Documents
table below. This document is a compilation of device and documentation errata,
specification clarifications and changes. It is intended for hardware system
manufacturers and software developers of applications, operating systems, or tools.
Information types defined in Nomenclature are consolidated into the specification
update and are no longer published in other documents.
This document may also contain information that was not previously published.
Errata are design defects or errors. These may cause the Intel® Xeon® Processor E5
v3 Product Family’s behavior to deviate from published specifications. Hardware and
software designed to be used with any given stepping must assume that all errata
documented for that stepping are present on all devices.
Specification changes are modifications to the current published specifications.
These changes will be incorporated in any new release of the specification.
Specification clarifications describe a specification in greater detail or further
highlight a specification’s impact to a complex design situation. These clarifications will
be incorporated in any new release of the specification.
Documentation changes include typos, errors, or omissions from the current
published specifications. These will be incorporated in any new release of the
specification.
S-Spec number is a five-digit code used to identify products. Products are
differentiated by their unique characteristics, such as core speed, L2 cache size,
package type, etc., as described in the processor identification information table. Read
all notes associated with each S-Spec number.
Document title
Document
number/location
330783
330784
Note:Errata remain in the specification update throughout the product’s lifecycle, or until a
particular stepping is no longer commercially available. Under these circumstances,
errata removed from the specification update are archived and available upon request.
Specification changes, specification clarifications and documentation changes are
removed from the specification update when the appropriate changes are made to the
appropriate product specification or user documentation (datasheets, manuals, and
so forth).
Intel® Xeon® Processor E5 v3 Product Family 5
Processor Specification Update
February 2016
Identification Information
GRP1 LINE1
GRP1 LINE2
GRP1 LINE3
GRP1 LINE4
Legend Mark Text (Production Mark)
GRP1 LINE1: i{M}{C}YY
GRP1 LINE2: SUB-BRAND PROC#
GRP1 LINE3: SSPEC SPEED
GRP1 LINE4: {FPO}{E4}
Component Identification via Programming Interface
The Intel® Xeon® Processor E5 v3 Product Family stepping can be identified by the
following register contents:
Reserved
31:2827:2019:1615:1413:1211:87:43:0
Notes:
1. The Extended Family, Bits [27:20] are used in conjunction with the Family Code, specified in Bits [11:8], to
indicate whether the processor belongs to the Intel® 386, Intel® 486, Pentium®, Pentium 4, or Intel®
Core™ Processor Family.
2. The Extended Model, Bits [19:16] in conjunction with the Model Number, specified in Bits [7:4], are used to
identify the model of the processor within the processor’s family.
3. The Family Code corresponds to Bits [11:8] of the EDX register after RESET, Bits [11:8] of the EAX register
after the CPUID instruction is executed with a 1 in the EAX register, and the generation field of the Device ID
register accessible through Boundary Scan.
4. The Model Number corresponds to Bits [7:4] of the EDX register after RESET, Bits [7:4] of the EAX register
after the CPUID instruction is executed with a 1 in the EAX register, and the model field of the Device ID
register accessible through Boundary Scan.
5. The Stepping ID in Bits [3:0] indicates the revision number of that model. See Table 1 for the processor
stepping ID number in the CPUID information.
6. Refer to the Intel® 64 and IA-32 Architectures Software Developer’s Manual documentation for additional
information.
1. Intel® Xeon® Processor E5-1600 v3 and E5-2600 v3 Product Families VID codes will change due to temperature and/or
current load changes in order to minimize power of the part. For specific voltages, refer to the latest Intel® Xeon® Processor
E5-1600/2400/2600/4600 v3 Product Families Datasheet, Volume 1, Electrical, #330783.
2. Refer to the latest revision of the following documents for information on processor specifications and features: Intel® Xeon®
Processor E5-1600/2400/2600/4600 v3 Product Families Datasheet, Volume 1, Electrical, #330783; Intel® Xeon® Processor
E5-1600/2400/2600/4600 v3 Product Families, Volume Two, Registers, #330784.
3. Refer to the latest Intel® Xeon® Processor E5-1600 / 2600 / 4600 v3 Product Families Thermal Mechanical Specification and
Design Guide, #330786 for information on processor operating temperature and thermal specifications.
4. This SKU does not support Intel® Hyper-Threading Technology.
5. This SKU does not support Intel® Turbo Boost Technology.
6. This SKU is intended for workstations only and uses workstation specific use conditions for reliability assumptions.
7. Intel® Turbo Boost Technology performance varies depending on hardware, software and overall system configuration.
1. Intel® Xeon® Processor E5-1600 v3 and E5-2600 v3 Product Families VID codes will change due to temperature and/or
current load changes in order to minimize power of the part. For specific voltages, refer to the latest Intel® Xeon® Processor
E5-1600/2400/2600/4600 v3 Product Families Datasheet, Volume 1, Electrical, #330783.
2. Refer to the latest revision of the following documents for information on processor specifications and features: Intel® Xeon®
Processor E5-1600/2400/2600/4600 v3 Product Families Datasheet, Volume 1, Electrical, #330783; Intel® Xeon® Processor
E5-1600/2400/2600/4600 v3 Product Families Datasheet, Volume Two, Registers, #330784.
3. Refer to the latest Intel® Xeon® Processor E5-1600 / 2600 / 4600 v3 Product Families Thermal Mechanical Specification and
Design Guide, #330786 for information on processor operating temperature and thermal specifications.
4. This SKU does not support Intel® Hyper-Threading Technology.
5. This SKU does not support Intel® Turbo Boost Technology.
6. This SKU is intended for workstations only and uses workstation specific use conditions for reliability assumptions.
7. Intel® Turbo Boost Technology performance varies depending on hardware, software and overall system configuration.
Model
Number
(W)
TDP
Intel® Turbo Boost Technology Maximum Core Frequency (GHz)
1. Intel® Xeon® Processor E5-1600 v3 and E5-2600 v3 Product Families VID codes will change due to temperature and/or
current load changes in order to minimize power of the part. For specific voltages, refer to the latest Intel® Xeon® Processor
E5-1600/2400/2600/4600 v3 Product Families Datasheet, Volume 1, Electrical, #330783.
2. Refer to the latest revision of the following documents for information on processor specifications and features: Intel® Xeon®
Processor E5-1600/2400/2600/4600 v3 Product Families Datasheet, Volume 1, Electrical, #330783; Intel® Xeon® Processor
E5-1600/2400/2600/4600 v3 Product Families Datasheet, Volume Two, Registers, #330784.
3. Refer to the latest Intel® Xeon® Processor E5-1600 / 2600 / 4600 v3 Product Families Thermal Mechanical Specification and
Design Guide, #330786 for information on processor operating temperature and thermal specifications.
4. This SKU does not support Intel® Hyper-Threading Technology.
5. This SKU does not support Intel® Turbo Boost Technology.
6. This SKU is intended for workstations only and uses workstation specific use conditions for reliability assumptions.
7. Intel® Turbo Boost Technology performance varies depending on hardware, software and overall system configuration.
Core
Frequency
(GHz)
Cores
#
Cache
Size
(MB)
Intel® AVX Turbo Boost Technology Maximum Core Frequency
Cores
1-2
Cores 3
Cores 4Cores 5Cores 6Cores 7Cores 8Cores
(MHz)
Notes
9+
12Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
Summary Tables of Changes
The following tables indicate the errata, specification changes, specification
clarifications, or documentation changes which apply to the Intel
v3 Product Family. Intel may fix some of the errata in a future stepping of the
component, and account for the other outstanding issues through documentation or
specification changes as noted. These tables uses the following notations:
Codes Used in Summary Tables
Stepping
X:Errata exists in the stepping indicated. Specification Change or
Clarification that applies to this stepping.
(No mark)
or (Blank box):This erratum is fixed in listed stepping or specification change
does not apply to listed stepping.
Page
(Page):Page location of item in this document.
Status
Doc:Document change or update will be implemented.
Plan Fix:This erratum may be fixed in a future stepping of the product.
Fixed:This erratum has been previously fixed.
No Fix:There are no plans to fix this erratum.
®
Xeon® Processor E5
Row
Change bar to left of table row indicates this erratum is either
new or modified from the previous version of the document.
Intel® Xeon® Processor E5 v3 Product Family 13
Processor Specification Update
February 2016
Table 4.Errata (Sheet 1 of 4)
Stepping
Number
HSE1XNo FixVM Exit May Set IA32_EFER.NXE When IA32_MISC_ENABLE Bit 34 is Set to 1
HSE75XNo FixThe System May Shut Down Unexpectedly During a Warm Reset.
HSE76XNo Fix
HSE77XNo FixPatrol Scrubbing of Mirrored Memory May Log Spurious Memory Errors
HSE78XNo FixMSR_TURBO_ACTIVATION_RATIO MSR Cannot be Locked
HSE79XNo FixDuplicate. Erratum removed
HSE80XNo FixAn APIC Timer Interrupt During Core C6 Entry May be Lost
HSE81XNo FixDDR4 CLK Signal May Incorrectly Float During Warm Reset or S3 State
HSE82XNo FixDDR4 Self Ref resh Entry May Result in Memory Read Errors
HSE83XNo Fix
HSE84XNo FixWrites To Some Control Register Bits Ignore Byte Enable
HSE85XNo Fix
HSE86XNo FixWarm Reset May Cause PCIe Hot-Plug Sequencing Failure
HSE87XNo FixPROCHOT# Assertion During Warm Reset May Cause Persistent Performance Reduction
HSE88XNo Fix
HSE89XNo FixPECI RdIAMSR Accesses During Boot Or Warm Reset May Cause The Processor to Hang
HSE90XNo FixDDR4 Rank Aliasing With Heavy Memory Traffic May Lead to a System Hang
HSE91XNo FixEarly Thermal Throttling May Occur
HSE92XNo FixIntel QPI Link Re-training After a Warm Reset or L1 Exit May be Unsuccessful
HSE93XNo FixPCIe* Header of a Malformed TLP is Logged Incorrectly
HSE94XNo FixAttempting to Enter ADR May Lead to Unpredictable System Behavior
HSE95XNo FixPECI Commands During Warm Reset May Lead to an Undefined Machine Check Error
HSE96XNo FixSpurious Corrected Errors May be Reported
HSE97XNo FixSystem Hang May Occur During Warm Reset Due to SMBUS Activity
HSE98XNo FixPECI PCS Package Identifier Command Operates Incorrectly on Certain SKUs
HSE99XNo FixIntel® Trusted Execution Technology Uses Incorrect TPM 2.0 NV Space Index Handles
HSE100XNo FixSurprise Down Error Status is Not Set Correctly on DMI Port
HSE101XNo FixPECI RdPkgConfig Command DRAM Services May Behave Incorrectly
HSE102XNo Fix
HSE103XNo Fix
HSE104XNo FixA P-State or C-State Transition May Lead to a System Hang
HSE105XNo FixAn IRET Instruction That Results in a Task Switch Does Not Serialize The Processor
HSE106XNo FixLoading a Microcode Update May Result in Higher Than Expected Idle Power
C-1,
M-1, R-2
StatusErrata
Accessing SB01BASE MMIO Space in The Presence of Bi-directional NTB Traffic May Result in
a System Hang
VT-d Memory Check Error on an Intel® QuickData Technology Channel May Cause All Other
Channels to Master Abort
Invalid Intel® QuickData Technology XOR Descriptor Source Addressing May Lead to
Unpredictable System Behavior
Operand-Size Override Prefix Causes 64-bit Operand Form of MOVBE Instruction to Cause a
Problem
Performance Monitoring OFFCORE_RESPOSE_{1,2} Events May Miscount
L3_MISS_REMOTE_HOP
Memory Power Consumption Reading May Not Be Accurate When Memory Channels Share a
VR
16Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
Table 4.Errata (Sheet 4 of 4)
Stepping
Number
HSE107XNo FixA Spurious Patrol Scrub Error May be Logged
HSE108XNo FixSome OFFCORE_RESPONSE Performance Monitoring Events May Undercount
HSE109XNo FixEnabling ADR with UDIMMs May Result in Unpredictable System Behavior
C-1,
M-1, R-2
StatusErrata
Table 5.Specification Clarifications
No.Specification clarifications
HSE1. TSX Instruction
1
Due to Erratum HSE44, TSX instructions are disabled and are only supported for software development. See your
Intel representative for details.
Table 6.Specification Changes
No.Specification changes
1None
Table 7.Documentation Changes
No.Documentation changes
1Datasheet Volume 3: Statement of Volatility
Intel® Xeon® Processor E5 v3 Product Family 17
Processor Specification Update
February 2016
Errata
HSE1VM Exit May Set IA32_EFER.NXE When IA32_MISC_ENABLE Bit 34 is
Set to 1
Problem:When “XD Bit Disable” in the IA32_MISC_ENABLE MSR (1A0H) bit 34 is set to 1, it
should not be possible to enable the “execute disable” feature by setting
IA32_EFER.NXE. Due to this erratum, a VM exit that occurs with the 1-setting of the
“load IA32_EFER” VM-exit control may set IA32_EFER.NXE even if IA32_MISC_ENABLE
bit 34 is set to 1. This erratum can occur only if IA32_MISC_ENABLE bit 34 was set by
guest software in VMX non-root operation.
Implication:Software in VMX root operation may execute with the “execute disable” feature enabled
despite the fact that the feature should be disabled by the IA32_MISC_ENABLE
MSR. Intel has not observed this erratum with any commercially available software.
Workaround: A virtual-machine monitor should not allow guest software to write to the
Status:For the affected steppings, see the Summary Tables of Changes.
HSE2Intel® QuickPath Interconnect (Intel® QPI) Layer May Report
Problem:Intel QPI may report an inband reset with no width change (error 0x22) correctable
Implication:An unexpected inband reset with no width change error may be logged.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
IA32_MISC_ENABLE MSR.
Spurious Correctable Errors
error upon exit from the L1 power state as logged in its IA32_MC{5, 20, 21}_STATUS
MSRs (415H,451H,455H).
this erratum.
HSE3NTB May Incorrectly Set MSI or MSI-X Interrupt Pending Bits
Problem:The NTB (Non-transparent Bridge) may incorrectly set MSI (Message Signaled
Interrupt) pending bits in MSIPENDING (BAR PB01BASE,SB01BASE; Offset 74H) while
operating in MSI-X mode or set MSI-X pending bits in PMSIXPBA (BAR PB01BASE,
SB01BASE; Offset 03000H) while operating in MSI mode.
Implication:Due to this erratum, NTB incorrectly sets MSI or MSI-X pending bits. The correct
pending bits are also set and it is safe to ignore the incorrectly set bits.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE4Processor May Issue Unexpected NAK DLLP Upon PCIe* L1 Exit
Problem:Upon exiting the L1 link power state, the processor’s PCIe port may unexpectedly issue
a NAK DLLP (Data Link Layer Packet).
Implication:PCIe endpoints may unexpectedly receive and log a NAK DLLP.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE5PECI DDR DIMM Digital Thermal Reading Returns Incorrect Value
Problem:When using the PECI RdPkgConfig() command to read PCS (Package Config Space)
Service 14 “DDR DIMM Digital Thermal Reading”, the value returned is incorrect.
Implication:Platform thermal management may not behave as expected.
18Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE6IIO CSR Lnkcon2 Field Selectable_De_Emphasis Cannot Be Set For
DMI2 Mode
Problem:The CSR lnkcon2 (Bus 0; Device 0; Function 0, Offset 0x1C0) field
selectable_de_emphasis (bit 6) cannot be set for a link when the DMI port is operating
at 5 GT/s.The documentation has the attribute of RW-O (read, write once), but the
processor incorrectly operates as read-only. This erratum does not occur when link is
operating as a PCIe port.
Implication:When the link is in DMI2 mode, the de-emphasis cannot be changed for an upstream
component.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE7PCIe* Receiver May Not Meet the Specification for AC Common Mode
Voltage And Jitter
Problem:Due to this erratum, PCIe receivers may not meet the specification for AC common
mode voltage (300 mV) and jitter (78.1 ps) at high temperatures when operating at 5
GT/s.
Implication:Specifications for PCIe receiver AC common mode voltage and jitter may not be met.
Intel has not observed this erratum on any commercially available system with any
commercially available PCIe devices.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE8Receiver Termination Impedance On PCIe* 3.0 Does Not Comply With
The Specification
Problem:The PCIe* Base Specification revision 3.0 defines ZRX-HIGH-IMP-DC-NEG and ZRX-
HIGH-IMP-DC-POS for termination impedance of the receiver. The specified impedance
for a negative voltage (-150 mV to 0V) is expected to be greater than 1 Kohm.
Sampled measurements of this impedance as low as 400 ohms have been seen. The
specified impedance for a positive voltage (> 200 mV) is greater than 20 Kohms.
Sampled measurements of this impedance as low as 14.6 Kohms have been seen.
Implication:Intel has not observed functional failures from this erratum on any commercially
available platforms using any commercially available PCIe device.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE9Platform Recovery After a Machine Check May Fail
Problem:While attempting platform recovery after a machine check (as indicated by CATERR#
signaled from the legacy socket), the original error condition may prevent normal
platform recovery which can lead to a second machine check. A remote processor
detecting a second Machine Check Event will hang immediately.
Implication:Due to this erratum, it is possible a system hang may be observed during a warm reset
caused by a CATERR#.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
Intel® Xeon® Processor E5 v3 Product Family 19
Processor Specification Update
February 2016
HSE10USB3 xHCI Not Compatible With MSIs
Problem:When the PCH xHCI (Extensible Host Controller Interface) is configured to use MSI
Implication:Due to this erratum, unpredictable system behavior may result.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
interrupts, a PCIe device number conflict between the processor and xHCI controller
may cause the interrupts be routed incorrectly.
this erratum.
HSE11A Memory Channel With More Than 4 Ranks May Lead to a System
Hang
Problem:A memory controller channel with more than 4 ranks and with TRR (Targeted Row
Implication:Due to this erratum, the system may hang.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
Refresh) enabled may fail leading to a system hang. This erratum only impacts memory
channels with three dual-rank DDR4 RDIMMs.
this erratum.
HSE12Writing R3QPI Performance Monitor Registers May Fail
Problem:Due to this erratum, attempting to write R3QPI performance monitor registers (Bus 0;
Device 11; Functions 1,2,5,6; Offset 0xA0-0xF7) may be unsuccessful.
Implication:A failed write to one or more R3QPI performance monitor registers is likely to yield
Workaround: Consecutively write the identified registers twice with the same value before
Status:For the affected steppings, see the Summary Tables of Changes.
incorrect performance events counts.
performance monitoring is globally enabled.
HSE13Enabling PPD May Cause Unpredictable System Behavior
Problem:Enabling memory PPD (Precharge Power Down) may cause unpredictable system
behavior.
Implication:This erratum may cause unpredictable system behavior.
Workaround: PPD is not supported. Use Active Power Down (APD) mode only.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE14CPUID Extended Topology Enumeration Leaf May Indicate an
Incorrect Number of Logical Processors
Problem:The Extended Topology Enumeration Leaf of CPUID (EAX = 0xB) may return an
incorrect value in EBX[15:0] for the core level type (ECX[15:8] = 2). In this instance,
the number of logical processors at the core level reported in EBX[15:0] should reflect
the configuration as shipped by Intel.
Implication:Software that uses the referenced CPUID function may not properly initialize all logical
processors in the system or correctly report the actual number of factory-configured
logical processors.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
20Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
HSE15Intel QPI Link Re-training After a Warm Reset or L1 Exit May be
Unsuccessful
Problem:After a warm reset or an L1 exit, the Intel® QPI (Intel QuickPath Interconnect) links
may not train successfully.
Implication:A failed Intel QPI link can lead to reduced system performance or an inoperable
system.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE16VCCIN VR Phase Shedding is Disabled
Problem:Due to this erratum, the processor does not direct the VCCIN VR (voltage regulator) to
Implication:Platform power consumption may exceed expected levels during deep package C-
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
shed phases during low power states.
states.
HSE17Quad-rank DDR4 LRDIMMs May Not be Properly Calibrated
Problem:Quad-rank LRDIMMs require calibration of all four of their DRAM ranks. Due to this
erratum, only half of the ranks are calibrated.
Implication:This erratum applies when a memory channel has three quad-rank DDR4 LRDIMMs.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
Uncalibrated ranks can result in higher correctable and uncorrectable error rates.
HSE18Possible Non-Optimal Electrical Margins on The DDR Command Bus
Problem:The processor periodically adjusts drive strength for DDR signals to optimize electrical
margins. Due to this erratum, the drive strength on the DDR command bus may be
incorrectly adjusted.
Implication:Reduced electrical margins on the command bus can lead to higher error rates possibly
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
affecting system stability.
this erratum.
HSE19PECI Commands During Reset May Result in Persistent Timeout
Response
Problem:Due to this erratum, a PECI (Platform Environment Control Interface) command other
Implication:Future PECI commands other than GetDIB(), Ping(), and GetTemp() will not be
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
than GetDIB(), Ping(), or GetTemp() received before RESET_N is de-asserted may
result in a timeout (0x81 completion code) for all subsequent such commands.
serviced after this erratum occurs.
HSE20System May Hang When Using the TPH Prefetch Hint
Problem:When all enabled cores on a socket are simultaneously in core C3, core C6, or package
C6 state and a PCIe* TPH (Transaction layer packet Processing Hint) with the prefetch
hint set is received, the system may hang.
Implication:Due to this erratum, the system may hang.
Intel® Xeon® Processor E5 v3 Product Family 21
Processor Specification Update
February 2016
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE21TS1s Do Not Convey The Correct Transmitter Equalization Values
During Recovery.RcvrLock
Problem:The PCIe* 3.1 Base Specification requires that TS1s sent during Recovery.RcvrLock
following 8.0 GT/s EQ (adaptive equalization) contain the final transmitter preset
number and coefficient values that were requested by an endpoint during phase 2 of
EQ. Due to this erratum, TS1s with incorrect transmitter preset number values may be
sent during Recovery.RcvrLock following 8.0 GT/s adaptive equalization.
Implication:Endpoints that check these values may, when unexpected values are found, request
equalization restart in subsequent TSs it sends. If EQ requests from the endpoint are
supported in the BIOS or OS, EQ will be restarted and the link may continue this EQ
loop indefinitely.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE22MSR_TEMPERATURE_TARGET MSR May Read as '0'
Problem:Due to this erratum, reading the MSR_TEMPERATURE_TARGET MSR (1A2H) may
incorrectly return '0'.
Implication:Software that depends on the contents of the MSR_TEMPERATURE_TARGET MSR may
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
not behave as expected.
HSE23PECI RdIAMSR() Command May Fail After Core C6 State is Entered
Problem:Reading core Machine Check Bank registers using the PECI (Platform Environment
Control Interface) RdIAMSR()command may fail after core C6 state has been entered.
Implication:Invalid data may be returned when using PECI to read core Machine Check Bank
registers.
Workaround: It is possible for the BIOS to contain a workaround for this erratum
Status:For the affected steppings, see the Summary Tables of Changes.
HSE24Processor May Not Enter PC3 or PC6 State
Problem:Due to this erratum, the processor may not enter PC3 (Package C3) or PC6 (Package
Implication:The processor may not meet power or thermal operating targets.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
C6) state when Intel Server Platform Services firmware is in use.
HSE25Disabling Intel® QPI L1 State May Cause a Machine Check On PC3 or
PC6 Exit
Problem:If Intel® QPI L1 state is disabled, exiting PC3 (Package C3) or PC6 (Package C6) state
Implication:Disabling Intel QPI link low power states can lead to a Machine Check if deep Package
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
may signal a Machine Check with IA32_MC4_STATUS_MSCOD (bits [31:16]) = 0x70XX.
C-states are enabled.
this erratum.
22Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
HSE26CLTT May Cause BIOS To Hang On a Subsequent Warm Reset
Problem:If CLTT (Closed Loop Thermal Throttling) is enabled when a warm reset is requested,
Implication:This erratum may lead to a BIOS hang. The warm reset request will fail, along with
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
due to this erratum, the processor will resume DIMM temperature polling before the
memory sub-system has been re-initialized.
subsequence warm reset attempts. The failing condition is cleared by a cold reset.
this erratum.
HSE27Systems May Experience Uncorrectable Errors When Using 2133 MHz
LRDIMM
Problem:Due to this erratum, a memory subsystem with 2133 MHz LRDIMMs may experience
uncorrectable memory errors.
Implication:The use of 2133 MHz LRDIMMs may lead to system failure.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE28PCIe* Extended Tag Field May be Improperly Set
Problem:The Extended Tag field in the TLP Header will not be zero for TLPs issued by PCIe ports
1a, 1b, 2c, 2d, 3c, and 3d even when the Extended Tag Field Enable bit in the Device
Control Register (Offset 08H, bit 8) is 0.
Implication:This does not affect ports 0, 2a, 2b, 3a and 3b. This will not result in any functional
issues when using device that properly track and return the full 8 bit Extended Tag
value with the affected ports. However, if the Extended Tag field is not returned by a
device connected to an affected port then this may result in unexpected completions
and completion timeouts.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE29A MOV to CR3 When EPT is Enabled May Lead to an Unexpected Page
Fault or an Incorrect Page Translation
Problem:If EPT (extended page tables) is enabled, a MOV to CR3 or VMFUNC may be followed by
an unexpected page fault or the use of an incorrect page translation.
Implication:Guest software may crash or experience unpredictable behavior as a result of this
erratum.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE30The System May Hang When a C/A Parity Error is Detected
Problem:Due to this erratum, detection of a C/A (Command/Address) parity error by the
Implication:System may experience a hang condition in the presence of C/A parity errors.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
Intel® Xeon® Processor E5 v3 Product Family 23
Processor Specification Update
February 2016
memory controller can lead to a system hang.
HSE31A C/A Parity Error When DDR4 is Operating at 2133 MHz May Result in
Unpredictable System Behavior
Problem:Due to this erratum, when DDR4 is operating at 2133MHz and a C/A (Command/
Address) parity error occurs while exiting a package C-state then unpredictable system
behavior may occur.
Implication:Due to this erratum, the system may experience unpredictable system behavior.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
this erratum.
HSE32Enabling Isochronous Transfers May Result in Unpredictable System
Behavior
Problem:Enabling isochronous transfers may lead to spurious correctable memory errors,
uncorrectable memory errors, patrol scrub errors and unpredictable system behavior.
Implication:The system may hang, report spurious memory errors, or behave unpredictably.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
this erratum.
HSE33Enabling Secondary To Primary Snoop Overrides On NTB May Cause a
Hang
Problem:Due to this erratum, NTB (Non-Transparent Bridge) completions may be dropped when
Secondary to Primary Snoop Overrides are enabled.
Implication:The system may hang or experience timeout machine checks when the secondary to
Workaround: None identified. A BIOS code change has been identified and may be implemented to
Status:For the affected steppings, see the Summary Tables of Changes.
primary snoop override is enabled. This erratum does not affect primary to secondary
snoop override.
avoid this erratum.
HSE34Memory Controller tsod_present Settings Being Improperly Cleared
Problem:On single Home Agent configurations, due to this erratum, the processor interferes with
TSOD (thermal sensor on DIMM) usage by incorrectly clearing the tsod_present field
(bits[7:0]) of the smbcntl_1 CSR (Bus 0; Device 19; Function 0; Offset 0x198) after
BIOS writes that field.
Implication:Closed Loop Thermal Throttle will not work as expected.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE35DDR4 Protocol May be Violated During C/A Parity Error Handling
Problem:Due to this erratum, if the error information in the DRAM is accessedbecause of a C/A
(Command/Address) parity error, DDR4 DIMM protocol rules may be violated leading to
unpredictable system behavior.
Implication:Attempts to access DDR4 DIMM error information may lead to unpredictable system
behavior.
Workaround: None identified. The en field (bit 31) of erf_ddr4_cmd_reg[4-0] CSRs (Bus 1; Device
Status:For the affected steppings, see the Summary Tables of Changes.
24Intel® Xeon® Processor E5 v3 Product Family
20,21,23,24; Function 0,1; Offset 0x24C, 0x250, 0x254, 0x258, 0x26C) must not be
set, preventing access to error information.
Processor Specification Update
February 2016
HSE36DDR4 Power Down Timing Violation
Problem:When DDR4 is operating at 2133 MHz, the processor’s memory control may violate the
JEDEC t
Implication:Violation of timing specifications can lead to unpredictable system behavior; however,
Intel has not observed this erratum to impact the operation of any commercially
available system using validated DIMMs by Intel Platform Memory Operations.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
PRPDEN timing specification.
HSE37Correctable Memory ECC Errors May Occur at Boot
Problem:With memory lockstep enabled, the system may experience correctable memory errors
during boot with IA32_MCi_STATUS.MCACOD= 0x009x (where x is 0,1,2, or 3 and
indicates the channel number reporting the error)
Implication:The system may experience correctable memory errors.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE38Remote P2P MCTP Transactions Cause Requests Timeouts
Problem:Remote P2P (Peer to Peer) MCTP (Management Component Transport Protocol)
messages may cause timeouts with IA32_MCi_STATUS.MSCOD=0x000c.
Implication:Due to this erratum, the system may hang.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE39Reserving Resources For Isochronous Transfers With Non-Posted
Prefetching Enabled May Cause a Hang
Problem:Resources in the IIO (Integrated I/O) unit are reserved for isochronous transfers to
ensure performance guarantees are met. Due to this erratum, enabling non-posted
prefetching in the IIO when resources are reserved for isochronous traffic may result in
a hang.
Implication:Due to this erratum, configuring the IIO unit to prefetch may result in a system hang.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE40BT Timeouts May Cause Spurious Machine Checks
Problem:The BT (Backup Tracker) timeout logic in the Home Agent can trigger spuriously,
causing false machine checks indicated by IA32_MCi_STATUS.MSCOD=0x0200.
Implication:Due to this erratum, timeout machine check may occur.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE41PCIe* Type 1 VDMs May be Silently Dropped
Problem:Due to this erratum, a PCIe Type 1 VDMs (Vendor Defined Message) is silently dropped
unless the vendor ID is the MCTP (Management Component Transport Protocol) value
of 0x1AB4.
Implication:PCIe Type 1 VDMs may be unexpectedly dropped. Intel has not observed this erratum
Workaround: None identified.
to impact the operation of any commercially available system.
Intel® Xeon® Processor E5 v3 Product Family 25
Processor Specification Update
February 2016
Status:For the affected steppings, see the Summary Tables of Changes.
HSE42Full Duplex NTB Traffic Can Cause a System Hang
Problem:If two PCIe* endpoints target traffic to PB23BASE (Bus 0; Device 3; Function 0; Offset
0x18, 0x1c) and PB45BASE (Bus 0; Device 3; Function 0; Offset 0x20, 0x24) registers
at the same time, a deadlock can result.
Implication:Due to this erratum, the system may hang.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
this erratum.
HSE43CONFIG_TDP_NOMINAL CSR Implemented at Incorrect Offset
Problem:The PCIe* Base Specification indicates that Configuration Space Headers have a base
address register at offset 0x10. Due to this erratum, the Power Control Unit’s
CONFIG_TDP_NOMINAL CSR (Bus 1; Device 30; Function 3; Offset 0x10) is located
where a base address register is expected.
Implication:Software may treat the CONFIG_TDP_NOMINAL CSR as a base address register leading
to a failure to boot.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE44Software Using Intel® TSX May Result in Unpredictable System
Behavior
Problem:Under a complex set of internal timing conditions and system events, software using
the Intel TSX (Transactional Synchronization Extensions) instructions may result in
unpredictable system behavior.
Implication:This erratum may result in unpredictable system behavior.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE45A Machine-Check Exception Due to Instruction Fetch May Be Delivered
Before an Instruction Breakpoint
Problem:Debug exceptions due to instruction breakpoints take priority over exceptions resulting
Implication:Instruction breakpoints may not operate as expected in the presence of a poisoned
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
from fetching an instruction. Due to this erratum, a machine-check exception resulting
from the fetch of an instruction may take priority over an instruction breakpoint if the
instruction crosses a 32-byte boundary and the second part of the instruction is in a
32-byte poisoned instruction fetch block.
instruction fetch block.
HSE46Spurious Corrected Errors May be Reported
Problem:Due to this erratum, spurious corrected errors may be logged in the IA32_MC0_Status
register with the valid field (bit 63) set, the uncorrected error field (bit 61) not set, a
Model Specific Error Code (bits [31:16]) of 0x000F, and an MCA Error Code (bits
[15:0]) of 0x0005. If CMCI is enabled, these spurious corrected errors also signal
interrupts.
Implication:When this erratum occurs, software may see corrected errors that are benign. These
corrected errors may be safely ignored.
Workaround: None identified.
26Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
Status:For the affected steppings, see the Summary Tables of Changes.
HSE47PCIe* LBMS Bit Incorrectly Set
Problem:If a PCIe Link autonomously changes width or speed for reasons other than to attempt
to correct unreliable Link operation, the Port should set LABS bit (Link Autonomous
Bandwidth Status) (Bus 0; Device 0; Function 0 and Device 1; Function 0-1 and Device
2-3; Function 0-3; Offset 0x1A2; bit 15). Due to this erratum, the processor will not set
this bit and will incorrectly set LBMS bit (Link Bandwidth Management Status) (Bus 0;
Device 0; Function 0 and Device 1; Function 0-1 and Device 2-3; Function 0-3; Offset
0x1A2; bit14) instead.
Implication:Software that uses the LBMS bit or LABS bit may behave incorrectly.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE48Power Consumed During Package C6 May Exceed Specification
Problem:Due to this erratum, the processor power usage may be higher than specified for the
Implication:Systems may experience increased power consumption while the processor is in
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
VCCIN and/or IIO domains while in Package C6 state.
Package C6.
this erratum.
HSE49Platform Performance Degradation When C1E is Enabled
Problem:Due to this erratum, when C1E is enabled and after the processor has entered Package
C1E state, core clock frequency becomes limited to its minimum value (sometimes
referred to as Pn) until the system exits Package C3 state (or deeper) or the system is
reset.
Implication:When this erratum occurs, operating frequency will be lower than expected. Note: After
a Package C3 exit, re-entering Package C1E state re-imposes this erratum’s frequency
limit.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE50Memory Power Down Entry May Lead to Unpredictable System
Behavior
Problem:Due to this erratum, the processor may violate DDR4 page close protocol at power
Implication:When this erratum occurs, it may result in unpredictable system behavior.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
down entry.
this erratum.
HSE51PCIe* Correctable Error Status Register May Not Log Receiver Error at
8.0 GT/s
Problem:Due to this erratum, correctable PCIe receiver errors may not be logged in the DPE field
Implication:Correctable receiver errors during 8.0 GT/s operation may not be visible to the OS or
Workaround: None identified.
Intel® Xeon® Processor E5 v3 Product Family 27
Processor Specification Update
February 2016
(bit 15) of the PCISTS CSR (Bus: 0; Device 1,2,3; Function 0-1, 0-3, 0-3; Offset 6H)
when operating at 8.0 GT/s.
driver software.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE52Sending PECI Messages Concurrently Over Two Interfaces May Result
in a System Hang
Problem:Sending messages concurrently via the PECI (Platform Environment Control Interface)
channel and the IBPECI (In-Band PECI) channel during Package C-State transitions
may result in a system hang.
Implication:Due to this erratum, the system may hang.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE53Platform Recovery After a Machine Check May Fail
Problem:While attempting platform recovery after a machine check (as indicated by CATERR#
signaled from the legacy socket), the original error condition may prevent normal
platform recovery which can lead to a second machine check. A remote processor
detecting a second Machine Check Event will hang immediately
Implication:Due to this erratum, it is possible a system hang may be observed during a warm reset
caused by a CATERR#.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE54Intel® Turbo Boost Technology Does Not Behave As Expected
Problem:Due to this erratum, the maximum turbo frequency may be incorrectly set to the
maximum non-turbo frequency after BIOS initialization completes.
Implication:Intel® Turbo Boost Technology may appear to be disabled.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE55PCIe* Hot Plug Slot Status Register May Not Indicate Command
Completed
Problem:The PCIe Base Specification requires a write to the Slot Control register (Offset A8H) to
generate a hot plug command when the downstream port is hot plug capable. Due to
this erratum, a hot plug command is generated only when one or more of the Slot
Control register bits [11:6] are changed.
Implication:Writes to the Slot Control register that leave bits [11:6] unchanged will not generate a
hot plug command and will therefore not generate a command completed event.
Software that expects a command completed event may not behave as expected.
Workaround: It is possible for software to implement a one-second timeout in lieu of receiving a
Status:For the affected steppings, see the Summary Tables of Changes.
command completed event.
HSE56Local PCIe* P2P Traffic on x4 Ports May Cause a System Hang
Problem:Under certain conditions, P2P (Peer-to-Peer) traffic with x4 PCIe ports on the same
processor (i.e., local) may cause a system hang.
Implication:Due to this erratum, the system may hang.
Workaround: None identified. Local P2P traffic should not be used to or from x4 PCIe ports.
Status:For the affected steppings, see the Summary Tables of Changes.
28Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
HSE57NTB Operating In NTB/RP Mode May Complete Transactions With
Incorrect ReqID
Problem:When the NTB (Non-Transparent Bridge) is operating in NTB/RP (NTB Root Port mode)
it is possible for transactions to be completed with the incorrect ReqID (Requester ID).
This erratum occurs when an outbound transaction is aborted before a completion for
inbound transaction is returned.
Implication:Due to this erratum, a completion timeout and an unexpected completion may be seen
by the processor connected to the NTB/RP. Intel has not observed this erratum with
any commercially available system.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE58Incorrect Single-Core Turbo Ratio Limit May be Applied When Using
AVX Instructions
Problem:The AVX single-core turbo ratio limit may not be applied correctly, unnecessarily
limiting the core frequency.
Implication:Single-core AVX workloads may not achieve single-core AVX turbo limit in some cases.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
This erratum does not apply to non-AVX workloads or multi-core workloads.
HSE59LLC Error Conditions May be Dropped or Incorrectly Signaled
Problem:When two LLC (last level cache) errors happen in close proximity, a UCNA
(uncorrectable no action required) machine check may be dropped or a spurious
machine check or CMCI (Corrected Machine Check Interrupt) may be issued. Further,
when this erratum occurs, the merged CBo LLC machine check bank IA32_MC[1719]_STATUS MSRs may be incorrect.
Implication:IA32_MC[17-19]_STATUS MSR may not reflect most current error.
Workaround: It is possible for the BIOS to contain a partial workaround for this erratum. The
Status:For the affected steppings, see the Summary Tables of Changes.
workaround does not address the potential dropped UCNA machine check.
HSE60Unexpected System Behavior May Occur Following Virtualization of
Some APIC Writes
Problem:If the "virtual-interrupt delivery" VM-execution control is enabled, unexpected system
behavior may occur following virtualization of the MOV to CR8 instruction, memorymapped accesses to the EOI (End of Interrupt), TPR (Task Priority Register), or the
interrupt command register to send an interprocessor interrupt to itself. This erratum
does not apply to the corresponding WRMSR access.
Implication:The unexpected system behavior may result in incorrect instruction execution, EPT
(Extended Page Table) violation, page fault, or similar event. These may cause
incorrect guest execution or may lead a virtual-machine monitor to terminate the
virtual machine that was running at the time.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE61A DDR4 C/A Parity Error in Lockstep Mode May Result in a Spurious
Uncorrectable Error
Problem:If a memory C/A (Command/Address) parity error occurs while the memory subsystem
Intel® Xeon® Processor E5 v3 Product Family 29
Processor Specification Update
February 2016
is configured in lockstep mode then the channel that observed the error will properly
log the error but the associated channel in lockstep will incorrectly log an uncorrectable
error in its IA32_MCi_STATUS MSR.
Implication:Due to this erratum, incorrect logging of an uncorrectable memory error in
IA32_MCi_STATUS may occur.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE62Some IMC and Intel QPI Functions Have Incorrect PCI CAPPTR Values
Problem:The PCI CAPPTR (Capability Pointer Register) is defined to contain the offset to the
capabilities list structure when the PCI PCISTS (PCI Status Register) bit 4
(Capabilities_List) is set to 1. Due to this erratum, CAPPTR (offset 0x34) should hold a
value of 0x40 but is instead zero for these IMC (Integrated Memory Controller) and
Intel® QPI (Intel QuickPath Interconnect) device:
Device 8, functions 3,5,6
Device 9, functions 3,5,6
Device 10, functions 3,5,6
Device 19, functions 0-5
Device 20, functions 0-3
Device 21, functions 0-3
Device 22, functions 0-3
Device 23, functions 0-3
Implication:Software that depends on CAPPTR to access additional capabilities may not behave as
expected.
Workaround: Software that needs to access these capabilities must take this erratum into account.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE63PCIe* TLP Translation Request Errors Are Not Properly Logged For
Invalid Memory Writes
Problem:A PCIe Memory Write TLP (Transaction Layer Packet) with an AT field value of 01b
(address translation request) does not set the UR (Unsupported Request) bit
(UNCERRSTS CSR, Bus 0; Device 0; Function 0; Offset 0x14C; Bit 20) as required by
the PCIe Base Specification.
Implication:System or software monitoring error status bits may not be notified of an unsupported
request. When this erratum occurs, the processor sets the
'advisory_non_fatal_error_status' bit (CORERRSTS CSR, Bus 0; Device 0; Function 0;
Offset 0x158; Bit 13) and drops the failing transaction.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE64The TSC May be Inaccurate When Per Core P-States Are Enabled
Problem:Due to this erratum, when per core P-States are enabled, the TSC (Time Stamp
Counter) may not be synchronized across the processor's cores.
Implication:The RDTSC (Read Time Stamp Counter) instruction may return a value that is not
monotonically increasing.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
30Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
HSE65Dual HA Processors With CLTT Enabled And no Memory on Channels 0
And 1 May Hang During Warm Reset
Problem:A dual HA (Home Agent) processor may hang during an attempted warm reset when
there is no memory installed on memory channels 0 and 1 and CLTT (Closed Loop
Thermal Throttling) is enabled.
Implication:Due to this erratum, a system may hang during warm reset. This erratum does not
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
occur if the processor has a single HA or at least one DIMM is installed on memory
channel 0 or 1.
HSE66PCIe* Slave Loopback May Transmit Incorrect Sync Headers
Problem:The PCIe Base Specification requires that, in the Loopback.Active state, a loopback
slave re-transmits the received bit stream bit-for-bit on the corresponding Tx. If the
link is directed to enter loopback slave mode at 8 GT/s via TS1 ordered sets with both
the Loopback and Compliance Receive bits set, the processor may place sync headers
in incorrect locations in the loopback bit stream.
Implication:In PCIe CEM (Card Electromechanical specification) Rx compliance testing directing the
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
link to loopback slave mode, the received data may not be correctly re-transmitted on
the Tx, causing the test to fail.
HSE67Consecutive PECI RdIAMSR Commands When Core C6 is Enabled May
Cause a System Hang
Problem:Consecutive PECI (Platform Environment Control Interface) RdIAMSR commands to
access core Machine Check MSRs can result in a system hang when core C6 state is
enabled.
Implication:When this erratum occurs, PECI commands can lead to a system hang.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE68Data Poisoning May Not Behave as Expected
Problem:When data poisoning is enabled, poisoned data consumption may not log and signal an
Implication:Due to this erratum, unpredictable system behavior may result.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
uncorrected machine check error.
HSE69Enabling TRR With DDR4 LRDIMMs May Lead to Unpredictable System
Behavior
Problem:Due to this erratum, TRR (Targeted Row Refresh) is not compatible with DDR4
Implication:Unpredictable system behavior may occur.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
Intel® Xeon® Processor E5 v3 Product Family 31
Processor Specification Update
February 2016
LRDIMMs.
this erratum.
HSE70Warm Reset May Cause PCIe Hot-Plug Sequencing Failure
Problem:The Integrated I/O unit uses the VPP (Virtual Pin Port) to communicate with power
Implication:During or shortly after a warm reset, when this erratum occurs, PCIe Hot-Plug
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
controllers, switches, and LEDs associated with PCIe Hot-Plug sequencing. Due to this
erratum, a warm reset occurring when a VPP transaction is in progress may result in an
extended VPP stall, termination of the in-flight VPP transaction, or a transient power
down of slots subject to VPP power control.
sequencing may experience transient or persistent failures or slots may experience
unexpected transient power down events. In certain instances, a cold reset may be
needed to fully restore operation.
HSE71Performance Monitor Instructions Retired Event May Not Count
Consistently
Problem:The Performance Monitor Instructions Retired event (Event C0H; Umask 00H) and the
instruction retired fixed counter IA32_FIXED_CTR0 MSR (309H) are used to count the
number of instructions retired. Due to this erratum, certain internal conditions may
cause the counter(s) to increment when no instruction has retired or to intermittently
not increment when instructions have retired.
Implication:A performance counter counting instructions retired may over count or under count.
The count may not be consistent between multiple executions of the same code.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE72C/A Parity Error Injection May Cause the System to Hang
Problem:When C/A (Command Address) parity error injections are occurring too frequently, the
home agent may be prevented from completing memory transactions. This may result
in an internal timer error indicated by IA32_MCi_STATUS. MSCOD=0x0080 and
IA32_MCi_STATUS. MCACOD=0x0400.
Implication:Due to this erratum, the system may hang.
Workaround: Ensure there is at least 30µs of delay between injections.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE73Excessive Fan Speed
Problem:Due to this erratum, fan speed control does not correctly account for the thermal
profile at elevated operating temperatures.
Implication:When this erratum occurs, fan speed is higher than required unnecessarily increasing
fan power consumption and fan noise.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
Problem:The processor’s PCIe links can be disabled by setting the Link Disable bit in the Link
Implication:The processor’s inability to reach the Package C6 state will result in increased idle
Workaround: None identified. For an optimal processor power configuration, unused PCIe ports
32Intel® Xeon® Processor E5 v3 Product Family
Control register (LNKCON.link_disable) (Bus 0; Device 0,1,2,3; Function 0,0-1,0-3,0-3;
Offset 0xA0; Bit 4) to 1. Due to this erratum, configuring one or more PCIe ports to be
disabled prevents the processor from entering package C6 state.
power consumption.
should remain enabled.
Processor Specification Update
February 2016
Status:For the affected steppings, see the Summary Tables of Changes.
HSE75The System May Shut Down Unexpectedly During a Warm Reset.
Problem:Certain complex internal timing conditions present when a warm reset is requested can
prevent the orderly completion of in-flight transactions. It is possible under these
conditions that the warm reset will fail and trigger a full system shutdown.
Implication:When this erratum occurs, the system will shut down and all machine check error logs
will be lost.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE76Accessing SB01BASE MMIO Space in The Presence of Bi-directional
NTB Traffic May Result in a System Hang
Problem:While transactions are originating from both sides of the NTB, accessing SB01BASE
MMIO space may cause the system to hang. For example, the NTB convention of using
SB01BASE MMIO doorbell or scratchpad registers to convey interrupt messages across
the NTB may result in a system hang. This erratum does not apply if transactions
originate from only one side of the NTB.
Implication:Due to this erratum, the system may hang.
Workaround: It is possible for the BIOS to contain a workaround for this erratum. Alternatively, Split-
Status:For the affected steppings, see the Summary Tables of Changes.
BAR mode can be used to send interrupts between NTB connected systems. Intel has
released a new version of its NTB driver that offers support for split-BAR mode.
HSE77Patrol Scrubbing of Mirrored Memory May Log Spurious Memory Errors
Problem:The Patrol scrubber, when mirroring is enabled, may incorrectly identify certain data
Implication:Spurious memory errors and poisoned data may be logged when mirroring is enabled.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
patterns as poison data or as memory errors.
this erratum.
HSE78MSR_TURBO_ACTIVATION_RATIO MSR Cannot be Locked
Problem:Setting the TURBO_ACTIVATION_RATIO_LOCK field (bit 31) of the
MSR_TURBO_ACTIVATION_RATIO MSR (64CH) has no effect; it does not block future
writes to the MSR_TURBO_ACTIVATION_RATIO MSR.
Implication:Software cannot rely on locking MSR_TURBO_ACTIVATION_RATIO MSR.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE79Duplicate. Erratum removed.
HSE80An APIC Timer Interrupt During Core C6 Entry May be Lost
Problem:Due to this erratum, an APIC timer interrupt coincident with the core entering C6 state
may be lost rather than held for servicing later.
Implication:A lost APIC timer interrupt may lead to missed deadlines or a system hang.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
Intel® Xeon® Processor E5 v3 Product Family 33
Processor Specification Update
February 2016
HSE81DDR4 CLK Signal May Incorrectly Float During Warm Reset or S3 State
Problem:The processor may not drive the DDR4 clock signal during warm reset or during S3
Implication:The anomalous electrical condition of an undriven clock signal can lead to correctable
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
State rather than driving the signal low as required by the JEDEC DDR4 specification.
and uncorrectable memory errors (IA32_MCi_STATUS.MCACOD = 0x009n, where n is
the channel number) after a warm reset or when resuming from S3 state.
HSE82DDR4 Self Refresh Entry May Result in Memory Read Errors
Problem:The processor may violate the JEDEC power down timing tCPDED parameter
specification associated with DDR4 self-refresh entry.
Implication:Correctable and/or uncorrectable memory read errors may occur.
Workaround: A BIOS code change has been identified and may be implemented as a workaround for
Status:For the affected steppings, see the Summary Tables of Changes.
this erratum.
HSE83VT-d Memory Check Error on an Intel® QuickData Technology Channel
May Cause All Other Channels to Master Abort
Problem:An Intel QuickData DMA access to Intel® VT-d protected memory that results in a
protected memory check error may cause master abort completions on all other Intel
QuickData DMA channels.
Implication:Due to this erratum, an error during Intel QuickData DMA access to an Intel® VT-d
protected memory address may cause a master abort on other Intel QuickData DMA
channels.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE84Writes To Some Control Register Bits Ignore Byte Enable
Problem:Due to this erratum, partial writes to some registers write the full register. The affected
Implication:Partial writes of the registers listed above may result in changes to register bytes that
Workaround: None identified. Use aligned, full-width DWORD (32-bit) read-modify-write sequencing
Status:For the affected steppings, see the Summary Tables of Changes.
registers are:
SADDBGMM2_CFG (Device 12; Function 0-7; Offset 0xA8 and Device 13; Function 0-6;
Offset 0xA8) and
LLCERRINJ_CFG (Device 12; Function 0-7; Offset 0xFC and Device 13; Function 0-6;
were intended to be unmodified.
to change a portion or portions of the registers listed.
Addressing May Lead to Unpredictable System Behavior
Problem:Intel QuickData Technology (i.e. Crystal Beach DMA v3.2) does not correctly halt and
report aborts on illegal source addresses placed in a CBDMA descriptor regardless of
type (Legacy or PQ). This abort condition may cause unpredictable system behavior.
Implication:This erratum may lead to unpredictable system behavior.
Status:For the affected steppings, see the Summary Tables of Changes.
34Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
HSE86Warm Reset May Cause PCIe Hot-Plug Sequencing Failure
Problem:The Integrated I/O unit uses the VPP (Virtual Pin Port) to communicate with power
Implication:During or shortly after a warm reset, when this erratum occurs, PCIe Hot-Plug
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
controllers, switches, and LEDs associated with PCIe Hot-Plug sequencing. Due to this
erratum, a warm reset occurring when a VPP transaction is in progress may result in an
extended VPP stall, termination of the inflight VPP transaction, or a transient power
down of slots subject to VPP power control.
sequencing may experience transient or persistent failures or slots may experience
unexpected transient power down events. In certain instances, a cold reset may be
needed to fully restore operation.
HSE87PROCHOT# Assertion During Warm Reset May Cause Persistent
Performance Reduction
Problem:Assertion of PROCHOT# after RESET# de-assertion but before BIOS has completed
reset initialization (indicated by CPL3) may result in persistent processor throttling.
Asserting PROCHOT# during and after RESET# assertion for FRB (Fault Resilient Boot)
tri-stating of the processor is not affected by this erratum.
Implication:When this erratum occurs, the resultant persistent throttling substantially reduces the
processor's performance.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE88Operand-Size Override Prefix Causes 64-bit Operand Form of MOVBE
Instruction to Cause a Problem
Problem:Execution of a 64-bit operand MOVBE instruction with an operand-size override
instruction prefix (66H) may incorrectly cause an invalid-opcode exception (#UD).
Implication:A MOVBE instruction with both REX.W=1 and a 66H prefix will unexpectedly cause an
Workaround: Do not use a 66H instruction prefix with a 64-bit operand MOVBE instruction.
Status:For the affected steppings, see the Summary Tables of Changes.
invalid-opcode exception (#UD). Intel has not observed this erratum with any
commercially available software.
HSE89PECI RdIAMSR Accesses During Boot Or Warm Reset May Cause The
Processor to Hang
Problem:The processor may hang when PECI RdIAMSR() accesses occur soon after processor
power-on or warm reset.
Implication:When this erratum occurs, the processor will hang and as a result PECI will be unable
to complete reads to the processor's machine check banks.
Workaround: All PECI RdIAMSR() accesses should be delayed until the CPU microcode update
Status:For the affected steppings, see the Summary Tables of Changes.
revision is non-zero. CPU microcode update revision can be accessed by PECI
RdPkgConfig() with index=0 and parameter=4.
HSE90DDR4 Rank Aliasing With Heavy Memory Traffic May Lead to a System
Hang
Problem:Under complex conditions, DDR4 rank aliasing during extended periods of heavy
memory traffic may lead to a system hang with IA32_MC{17-19}_STATUS.MSCOD =
0x000C.
Implication:DDR4 rank aliasing may affect the reliability of a highly utilized system.
Intel® Xeon® Processor E5 v3 Product Family 35
Processor Specification Update
February 2016
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE91Early Thermal Throttling May Occur
Problem:Systems engineered to published thermal specifications and requirements may throttle
early due to insufficient thermal margin on the processor.
Implication:When this erratum occurs, the thermal throttling may begin as much as 2°C early.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE92Intel QPI Link Re-training After a Warm Reset or L1 Exit May be
Unsuccessful
Problem:After a warm reset or an L1 exit, the Intel® QPI (Intel QuickPath Interconnect) links
may not train successfully.
Implication:A failed Intel QPI link can lead to reduced system performance or an inoperable
Workaround: It is possible for BIOS to contain processor configuration data and code changes as a
Status:For the affected steppings, see the Summary Tables of Changes.
system.
workaround for this erratum.
HSE93PCIe* Header of a Malformed TLP is Logged Incorrectly
Problem:If a PCIe port receives a malformed TLP (Transaction Layer Packet), an error is logged
in the UNCERRSTS register (Device 0; Function 0; Offset 14CH and Device 2-3;
Function 0-3; Offset 14CH). Due to this erratum, the header of the malformed TLP is
logged incorrectly in the HDRLOG register (Device 0; Function 0; Offset 164H and
Device 2-3; Function 0-3; Offset 164H).
Implication:The PCIe header of a malformed TLP is not logged correctly.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE94Attempting to Enter ADR May Lead to Unpredictable System Behavior
Problem:Due to this erratum, an attempt to transition the memory subsystem to ADR
(Asynchronous DRAM Self Refresh) mode may fail.
Implication:This erratum may lead to unpredictable system behavior.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE95PECI Commands During Warm Reset May Lead to an Undefined
Machine Check Error
Problem:After a warm reset request is made, generally by BIOS or in response to a processor
error, the processor is not able to accept PECI commands until the subsequent RESET#
signal assertion ends. A PECI command during this interval will log an undefined model
specific error code of 0x26 in bits[31:24] of the IA32_MCi_STATUS MSR (411H).
Implication:Due to this erratum, attempts to issue PECI commands to the processor while RESET#
is asserted may result in an unexpected error.
Workaround: An external agent that is able to issue PECI commands must not issue PECI commands
Status:For the affected steppings, see the Summary Tables of Changes.
36Intel® Xeon® Processor E5 v3 Product Family
during the interval from warm reset request to RESET# deassertion.
Processor Specification Update
February 2016
HSE96Spurious Corrected Errors May be Reported
Problem:Due this erratum, spurious corrected errors may be logged in the IA32_MC0_STATUS
Implication:When this erratum occurs, software may see corrected errors that are benign. These
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
register with the valid field (bit 63) set, the uncorrected error field (bit 61) not set, a
Model Specific Error Code (bits [31:16]) of 0x000F, and an MCA Error Code (bits
[15:0]) of 0x0005. If CMCI is enabled, these spurious corrected errors also signal
interrupts.
corrected errors may be safely ignored.
HSE97System Hang May Occur During Warm Reset Due to SMBUS Activity
Problem:SMBUS activity during warm reset may lead to a system hang.
Implication:Due to this erratum, a system hang may occur during warm reset if activity on the
SMBUS is present.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE98PECI PCS Package Identifier Command Operates Incorrectly on
Certain SKUs
Problem:The Max Thread ID value returned via PECI PCS (Package Config Space) command,
Package Identifier Service (Index 00), Parameter Value 0x0003, may be incorrect on
E5-2643V3 and E5-2637V3 SKUs.
Implication:A PECI agent may be given an incorrect count of logical processors.
Workaround: It is possible for BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
2.0 draft specification handles (indices) AUX 01800003, PS 01800001, and PO
01400003. Those handles conflict with the released TCG (Trusted Computing Group)
“Registry of reserved TPM 2.0 handles and localities”, version 1.0, revision 1.
Implication:Intel TXT TPM 2.0 handles may conflict with platform manufacturer or owner usage of
TPM NV space. Intel has not identified any functional impact due to this erratum.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE100Surprise Down Error Status is Not Set Correctly on DMI Port
Problem:Due to this erratum, the Surprise_down_error_status (UNCERRSTS Device 0; Function
Implication:Surprise down errors will not be logged for the DMI port. Software that relies on this
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
0; Offset 0x14C; bit 5) is not set to 1 when DMI port detects a surprise down error.
status bit may not behave as expected.
HSE101PECI RdPkgConfig Command DRAM Services May Behave Incorrectly
Problem:The PECI (Platform Environment Control Interface) RdPkgConfig command may return
incorrect results when accessing the DRAM Thermal Interface (indices 14 and 22).
Implication:Thermal monitoring and control using PECI may not behave as expected.
Intel® Xeon® Processor E5 v3 Product Family 37
Processor Specification Update
February 2016
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE102Performance Monitoring OFFCORE_RESPOSE_{1,2} Events May
Miscount L3_MISS_REMOTE_HOP
Problem:When a Performance Monitoring counter is configured to count
OFF_CORE_RESPONSE_{1,2} (Events B7H and B8H), data obtained for remote DRAM
may be attributed to L3_MISS_REMOTE_HOP0 (as programmed by
MSR_OFFCORE_RSP_{1,2} (MSRs 1A6H, 1A7H) bit 27) instead of
L3_MISS_REMOTE_HOP1 (bit 28) or L3_MISS_REMOTE_HOP2P (bit 29). Data provided
from remote caching agent associated with remote DRAM is unaffected.
Implication:L3_MISS_REMOTE_HOP0 may over count, while L3_MISS_REMOTE_HOP1 and
Workaround: None identified. Set all three configuration bits (L3_MISS_REMOTE_HOP0,
Status:For the affected steppings, see the Summary Tables of Changes.
L3_MISS_REMOTE_HOP2P may undercount.
L3_MISS_REMOTE_HOP1, L3_MISS_REMOTE_HOP2P) to obtain the total count of data
supplied by remote agents.
HSE103Memory Power Consumption Reading May Not Be Accurate When
Memory Channels Share a VR
Problem:When a single VR (voltage regulator) is used by more than one populated memory
Implication:RAPL (running average power limit) memory power regulation may not behave as
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
channel, power readings of these memory channels may not be accurate.
expected.
HSE104A P-State or C-State Transition May Lead to a System Hang
Problem:For a small subset of parts under elevated die temperature conditions, a P-state or C-
state transition may result in a system timeout or system shutdown.
Implication:When this erratum occurs, the system may shutdown or report a timeout error; Intel
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
has observed transaction completion timeouts and other internal timeouts.
HSE105An IRET Instruction That Results in a Task Switch Does Not Serialize
The Processor
Problem:An IRET instruction that results in a task switch by returning from a nested task does
Implication:Software which depends on the serialization property of IRET during task switching
Workaround: None identified. Software can execute an MFENCE instruction immediately prior to the
Status:For the affected steppings, see the Summary Tables of Changes.
not serialize the processor (contrary to the Software Developer's Manual Vol. 3 section
titled “Serializing Instructions”).
may not behave as expected. Intel has not observed this erratum to impact the
operation of any commercially available software.
IRET instruction if serialization is needed.
HSE106Loading a Microcode Update May Result in Higher Than Expected Idle
Power
Problem:A microcode update loaded by the operating system or virtual machine monitor may
change the power management configuration previously established by the BIOS.
38Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
Implication:Loading a microcode update after BIOS initialization completes may cause higher than
expected idle power.
Workaround: It is possible for the BIOS to contain a workaround for this erratum.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE107A Spurious Patrol Scrub Error May be Logged
Problem:When a memory ECC error occurs, a spurious patrol scrub error may also be logged on
another memory channel.
Implication:A patrol scrub correctable error may be incorrectly logged.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE108Some OFFCORE_RESPONSE Performance Monitoring Events May
Undercount
Problem:The performance monitoring events OFFCORE_RESPONSE (Events B7H and BBH)
should count uncore responses matching the request-response configuration specified
in MSR_OFFCORE_RSPs (1A6H and 1A7H, respectively) for core-originated requests.
However due to this erratum, COREWB (bit 3), PF_L3_DATA_RD (bit 7), PF_L3_RFO
(bit 8), PR_L3_CODE_RD (bit 9), SPLIT_LOCK_UC_LOCK (bit 10), and
STREAMING_STORES (bit 15) request types may undercount.
Implication:These performance monitoring events may not produce reliable results for the listed
request types.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
HSE109Enabling ADR with UDIMMs May Result in Unpredictable System
Behavior
Problem:When ADR (Asynchronous DIMM self-Refresh) is enabled during an S3 resume, the
processor may cause the UDIMM to prematurely exit self refresh.
Implication:Due to this erratum, unpredictable system behavior may occur when ADR is enabled.
Workaround: None identified.
Status:For the affected steppings, see the Summary Tables of Changes.
Intel® Xeon® Processor E5 v3 Product Family 39
Processor Specification Update
February 2016
Specification Changes
1.There are no specification changes in this specification update
revision.
§
40Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
Specification Clarifications
1.There are no specification clarifications in this specification update
revision.
§
Intel® Xeon® Processor E5 v3 Product Family 41
Processor Specification Update
February 2016
Documentation Changes
1.Datasheet Volume 3: Statement of Volatility
No Intel® Xeon® Processor E5 v3 Family processors retain any end user data
when powered down and/or when the parts are physically removed from the
socket.
§
42Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
Mixed Processors Within DP
Platforms
Mixed Processor Consistency Requirements
Intel supports dual processor (DP) configurations consisting of processors:
1. From the same power rating.
2. That support the same maximum Intel® QuickPath Interconnect (Intel® QPI) and
DDR4 memory speeds.
3. That share symmetry across physical packages with respect to the number of
logical processors per package, number of cores per package, number of Intel QPI
Interfaces, and cache topology.
4. That have identical Extended Family, Extended Model, Processor Type, Family Code
and Model Number as indicated by the function 1 of the CPUID instruction.
Note:Processors must operate with the same Intel QPI, DDR4 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 E5 v3 product family attributes, the following population rules and stepping
matrix have been developed to clearly define supported configurations.
1. Processors must be of the same power rating. For example, mixing of 95W Thermal
Design Power (TDP) processors is supported. Mixing of dissimilar TDPs in the same
platform is not supported (for example, 95W with 130W, and so forth).
2. Processors must operate at the same core frequency. Note: Processors within the
same power-optimization segment supporting different maximum core frequencies
(for example, a 2.93 GHz / 95 W and 2.66 GHz / 95W) can be operated within a
system. However, both must operated 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.
3. Processors must share symmetry across physical packages with respect to the
number of logical processors per package, number of Intel QPI Interfaces, and
cache topology.
4. Mixing dissimilar 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 model (as per CPUID instruction) is supported.
Details regarding the CPUID instruction are provided in the AP-487, Intel®
Processor Identification and the CPUID Instruction application note and Intel® 64
and IA-32 Architectures Software Developer’s Manual, Volume 2A.
5. After AND’ing the feature flag and extended feature flag 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 processors with the numerically
lower CPUID should be selected as the BSP.
6. The workarounds identified in this, and subsequent specification updates, must
properly applied to each processor in the system. Certain errata are specific to the
multiprocessor environment. Errata for all processor steppings will affect system
performance if not properly worked around.
Intel® Xeon® Processor E5 v3 Product Family 43
Processor Specification Update
February 2016
Mixed Steppings
Mixing processors of different steppings but the same model (as per CPUID instruction)
is supported provided there is no more than one stepping delta between the
processors, for example, S and S+1.
S and S+1 is defined as mixing of two CPU steppings in the same platform where one
CPU is S (stepping) = CPUID.(EAX=01h):EAX[3:0], and the other is S+1 =
CPUID.(EAX=01h):EAX[3:0]+1. The stepping ID is found in EAX[3:0] after executing
the CPUID instruction with Function 01h.
§
44Intel® Xeon® Processor E5 v3 Product Family
Processor Specification Update
February 2016
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.