Intel BX80644E52603V3 User Manual

Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

Reference Number: 330785-010US

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.

2 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

Contents

Revision History ........................................................................................................ 4

Preface ...................................................................................................................... 5

Affected Documents.............................................................................................5

Nomenclature .....................................................................................................5

Identification Information.........................................................................................6

Component Identification via Programming Interface ...............................................6

Component Marking Information ...........................................................................6

Summary Tables of Changes.................................................................................... 13

Codes Used in Summary Tables........................................................................... 13

Errata ...................................................................................................................... 18

Specification Changes.............................................................................................. 40

Specification Clarifications ...................................................................................... 41

Documentation Changes .......................................................................................... 42

Mixed Processors Within DP Platforms .................................................................... 43

Mixed Processor Consistency Requirements........................................................... 43

Intel® Xeon® Processor E5 v3 Product Family 3 Processor Specification Update February 2016

Revision History

Date Revision Description

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

4 Intel® 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.

Affected Documents

Intel® Xeon® Processor E5-1600/2400/2600/4600 v3 Product Families Datasheet Volume 1: Electrical

Intel® Xeon® Processor E5-1600/2400/2600/4600 v3 Product Families Datasheet Volume 2: Registers

Nomenclature

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

31:28 27:20 19:16 15:14 13:12 11:8 7:4 3:0

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.

Extended

Family

00000000b 0011b 00b 0110b 1111b C1=0002

00000000b 0011b 00b 0110b 1111b M1=0002

00000000b 0011b 00b 0110b 1111b R2=002

Extended

Model

Reserved

Processor

Type

Family

Code

Model

Number

Stepping

Component Marking Information

Figure 1. Intel® Xeon® Processor E5 v3 Product Families Top-side Markings (Example)

6 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

Table 1. Intel® Xeon® Processor E5-1600, E5-2600 and E5-4600 v3 Product Families

Identification (Sheet 1 of 2)

S-Spec

SR1XG C1 E5-2695 v3 0x306F2 2.3/DDR4 2133/9.6 120 14 35 2 1,2,3,7

SR1XN M1 E5-2690 v3 0x306F2 2.6/DDR4 2133/9.6 135 12 30 2 1,2,3,7

SR1XP M1 E5-2680 v3 0x306F2 2.5/DDR4 2133/9.6 120 12 30 2 1,2,3,7

SR1XR M1 E5-2660 v3 0x306F2 2.6/DDR4 2133/9.6 105 10 25 2 1,2,3,7

SR1XS M1 E5-2670 v3 0x306F2 2.3/DDR4 2133/9.6 120 12 30 2 1,2,3,7

SR1XV M1 E5-2658 v3 0x306F2 2.2/DDR4 2133/9.6 105 12 30 2 1,2,3,7

SR1XW M1 E5-2648L v3 0x306F2 1.8/DDR4 2133/9.6 75 12 30 2 1,2,3,7

SR1XZ M1 E5-2628L v3 0x306F2 2/DDR4 1866/8 75 10 25 2 1,2,3,7

SR20H R2 E5-1680 v3 0x306F2 3.2/DDR4 2133/NA 140 8 20 1 1,2,3,6,7

SR20J R2 E5-1650 v3 0x306F2 3.5/DDR4 2133/NA 140 6 15 1 1,2,3,6,7

SR20K R2 E5-1603 v3 0x306F2 2.8/DDR4 1866/NA 140 4 10 1 1,2,3,4,5,6

SR1Y1 M1 E5-2650L v3 0x306F2 1.8/DDR4 2133/9.6 65 12 30 2 1,2,3,7

SR1XH C1 E5-2683 v3 0x306F2 2/DDR4 2133/9.6 120 14 35 2 1,2,3,7

SR1XE C1 E5-2698 v3 0x306F2 2.3/DDR4 2133/9.6 135 16 40 2 1,2,3,7

SR200 R2 E5-2618L v3 0x306F2 2.3/DDR4 1866/8 75 8 20 1 1,2,3,7

SR1Y6 M1 E5-2687W v3 0x306F2 3.1/DDR4 2133/9.6 160 10 25 2 1,2,3,7

SR20L R2 E5-1630 v3 0x306F2 3.7/DDR4 2133/NA 140 4 10 1 1,2,3,6,7

SR1Y9 M1 E5-2685 v3 0x306F2 2.6/DDR4 2133/9.6 120 12 30 2 1,2,3,4,7

SR1YA M1 E5-2650 v3 0x306F2 2.3/DDR4 2133/9.6 105 10 25 2 1,2,3,7

SR202 R2 E5-2637 v3 0x306F2 3.5/DDR4 2133/9.6 135 4 15 1 1,2,3,7

SR203 R2 E5-2667 v3 0x306F2 3.2/DDR4 2133/9.6 135 8 20 1 1,2,3,7

SR204 R2 E5-2643 v3 0x306F2 3.4/DDR4 2133/9.6 135 6 20 1 1,2,3,7

SR20M R2 E5-1607 v3 0x306F2 3.1/DDR4 1866/NA 140 4 10 1 1,2,3,4,5,6

SR1XD C1 E5-2699 v3 0x306F2 2.3/DDR4 2133/9.6 145 18 45 2 1,2,3,7

SR1XF C1 E5-2697 v3 0x306F2 2.6/DDR4 2133/9.6 145 14 35 2 1,2,3,7

SR205 R2 E5-2640 v3 0x306F2 2.6/DDR4 1866/8 90 8 20 1 1,2,3,7

SR206 R2 E5-2630 v3 0x306F2 2.4/DDR4 1866/8 85 8 20 1 1,2,3,7

SR207 R2 E5-2620 v3 0x306F2 2.4/DDR4 1866/8 85 6 15 1 1,2,3,7

SR208 R2 E5-2623 v3 0x306F2 3/DDR4 1866/8 105 4 10 1 1,2,3,7

SR209 R2 E5-2630L v3 0x306F2 1.8/DDR4 1866/8 55 8 20 1 1,2,3,7

SR20A R2 E5-2603 v3 0x306F2 1.6/DDR4 1600/6.4 85 6 15 1 1,2,3,4,5

SR1YC M1 E5-2609 v3 0x306F2 1.9/DDR4 1600/6.4 85 6 15 2 1,2,3,4,5

SR20N R2 E5-1660 v3 0x306F2 3/DDR4 2133/NA 140 8 20 1 1,2,3,6,7

Stepping

Model

Number

CPUID

Core frequency (GHz)/Memory

(MHz)/

Intel QPI (GHz)

TDP (W)# Cores

Cache

size

(MB)

Home

Agents

Notes

Intel® Xeon® Processor E5 v3 Product Family 7 Processor Specification Update February 2016

Table 1. Intel® Xeon® Processor E5-1600, E5-2600 and E5-4600 v3 Product Families

Identification (Sheet 2 of 2)

S-Spec

SR20B R2 E5-2608L v3 0x306F2 2/DDR4 1866/6.4 50 6 15 1 1,2,3,7

SR20P R2 E5-1620 v3 0x306F2 3.5/DDR4 2133/NA 140 4 10 1 1,2,3,6,7

SR21P R2 E5-2608L v3 0x306F2 2/DDR4 1866/6.4 52 6 15 1 1,2,3,7

SR22P C1 E5-4660 v3 0x306F2 2.1/DDR4 2133/9.6 120 14 35 2 1,2,3,7,

SR22J C1 E5-4650 v3 0x306F2 2.1/DDR4 2133/9.6 105 12 30 2 1,2,3,7,

SR22L C1 E5-4640 v3 0x306F2 1.9/DDR4 1866/8 105 12 30 2 1,2,3,7,

SR22K C1 E5-4620 v3 0x306F2 2/DDR4 1866/8 105 10 25 2 1,2,3,7,

SR22S C1 E5-4610 v3 0x306F2 1.7/DDR4 1600/6.4 105 10 25 2 1,2,3,7,

SR22M C1 E5-4669 v3 0x306F2 2.1/DDR4 2133/9.6 135 18 45 2 1,2,3,7,

SR22N C1 E5-4667 v3 0x306F2 2/DDR4 2133/9.6 135 16 40 2 1,2,3,7,

SR22R M1 E5-4655 v3 0x306F2 2.9/DDR4 2133/9.6 135 6 30 2 1,2,3,7,

SR26R C1 E5-4648 v3 0x306F2 1.7/DDR4 1866/8 105 12 30 2 1,2,3,7,

SR22Q M1 E5-4627 v3 0x306F2 2.6/DDR4 2133/8 135 10 25 2 1,2,3,4,7,

SR233 M1 E5-2438L v3 0x306F2 1.8/DDR3 0/8 70 10 25 1 1,2,3,7,

SR235 R2 E5-2428L v3 0x306F2 1.8/DDR3 0/8 55 8 20 1 1,2,3,7,

SR234 M1 E5-1428L v3 0x306F2 2/DDR3 0/0 65 8 20 1 1,2,3,7,

SR232 R2 E5-2418L v3 0x306F2 2/DDR3 0/6.4 50 6 15 1 1,2,3,7,

SR236 R2 E5-2408L v3 0x306F2 1.8/DDR3 0/6.4 45 4 10 1 1,2,3,7,

Stepping

Model

Number

CPUID

Core frequency (GHz)/Memory

(MHz)/

Intel QPI (GHz)

TDP (W)# Cores

Cache

size

(MB)

Home

Agents

Notes

Notes:

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.

Table 2. Intel® Xeon® Processor E5-1600, E5-2600 and E5-4600 v3 Product Families

Turbo Bins (Sheet 1 of 3)

S-Spec

Stepping

SR1XG C1 E5-2695 v3 120 14 3.3 3.1 3 2.9 2.8 2.8 2.8 2.8 2.8 2.8 1,2,3,7

SR1XN M1 E5-2690 v3 135 12 3.5 3.3 3.2 3.1 3.1 3.1 3.1 3.1 3.1 3.1 1,2,3,7

SR1XP M1 E5-2680 v3 120 12 3.3 3.1 3 2.9 2.9 2.9 2.9 2.9 2.9 2.9 1,2,3,7

8 Intel® Xeon® Processor E5 v3 Product Family

Model

Number

(W)

TDP

Intel® Turbo Boost Technology Maximum Core Frequency (GHz)

Core

# Cores

Core 3Core4Core5Core6Core7Core8Core9Core10Core

1 -2

Processor Specification Update

Notes

11+

February 2016

Table 2. Intel® Xeon® Processor E5-1600, E5-2600 and E5-4600 v3 Product Families

Turbo Bins (Sheet 2 of 3)

S-Spec

Stepping

SR1XR M1 E5-2660 v3 105 10 3.3 3.1 3 2.9 2.9 2.9 2.9 2.9 2.9 NA 1,2,3,7

SR1XS M1 E5-2670 v3 120 12 3.1 2.9 2.8 2.7 2.6 2.6 2.6 2.6 2.6 2.6 1,2,3,7

SR1XV M1 E5-2658 v3 105 12 2.9 2.7 2.6 2.5 2.5 2.5 2.5 2.5 2.5 2.5 1,2,3,7

SR1XW M1 E5-2648L v3 75 12 2.5 2.3 2.2 2.1 2.1 2.1 2.1 2.1 2.1 2.1 1,2,3,7

SR1XZ M1 E5-2628L v3 75 10 2.5 2.3 2.2 2.2 2.2 2.2 2.2 2.2 2.2 NA 1,2,3,7

SR20H R2 E5-1680 v3 140 8 3.8 3.6 3.5 3.5 3.5 3.5 3.5 NA NA NA 1,2,3,6,7

SR20J R2 E5-1650 v3 140 6 3.8 3.6 3.6 3.6 3.6 NA NA NA NA NA 1,2,3,6,7

SR20K R2 E5-1603 v3 140 4 2.8 2.8 2.8 NA NA NA NA NA NA NA 1,2,3,4,5,6

SR1Y1 M1 E5-2650L v3 65 12 2.5 2.3 2.2 2.1 2.1 2.1 2.1 2.1 2.1 2.1 1,2,3,7

SR1XH C1 E5-2683 v3 120 14 3 2.8 2.7 2.6 2.5 2.5 2.5 2.5 2.5 2.5 1,2,3,7

SR1XE C1 E5-2698 v3 135 16 3.6 3.4 3.3 3.2 3.1 3 2.9 2.8 2.8 2.8 1,2,3,7

SR200 R2 E5-2618L v3 75 8 3.4 3 2.9 2.8 2.7 2.6 2.5 NA NA NA 1,2,3,7

SR1Y6 M1 E5-2687W v3 160 10 3.5 3.3 3.2 3.2 3.2 3.2 3.2 3.2 3.2 NA 1,2,3,7

SR20L R2 E5-1630 v3 140 4 3.8 3.8 3.8 NA NA NA NA NA NA NA 1,2,3,6,7

SR1Y9 M1 E5-2685 v3 120 12 3.3 3.1 3 2.9 2.8 2.8 2.8 2.8 2.8 2.8 1,2,3,4,7

SR1YA M1 E5-2650 v3 105 10 3 2.8 2.7 2.6 2.6 2.6 2.6 2.6 2.6 NA 1,2,3,7

SR202 R2 E5-2637 v3 135 4 3.7 3.6 3.6 NA NA NA NA NA NA NA 1,2,3,7

SR203 R2 E5-2667 v3 135 8 3.6 3.4 3.4 3.4 3.4 3.4 3.4 NA NA NA 1,2,3,7

SR204 R2 E5-2643 v3 135 6 3.7 3.6 3.6 3.6 3.6 NA NA NA NA NA 1,2,3,7

SR20M R2 E5-1607 v3 140 4 3.1 3.1 3.1 NA NA NA NA NA NA NA 1,2,3,4,5,6

SR1XD C1 E5-2699 v3 145 18 3.6 3.4 3.3 3.2 3.1 3 2.9 2.8 2.8 2.8 1,2,3,7

SR1XF C1 E5-2697 v3 145 14 3.6 3.4 3.3 3.2 3.1 3.1 3.1 3.1 3.1 3.1 1,2,3,7

SR205 R2 E5-2640 v3 90 8 3.4 3.2 3.1 3 2.9 2.8 2.8 NA NA NA 1,2,3,7

SR206 R2 E5-2630 v3 85 8 3.2 3 2.9 2.8 2.7 2.6 2.6 NA NA NA 1,2,3,7

SR207 R2 E5-2620 v3 85 6 3.2 2.9 2.8 2.7 2.6 NA NA NA NA NA 1,2,3,7

SR208 R2 E5-2623 v3 105 4 3.5 3.3 3.3 NA NA NA NA NA NA NA 1,2,3,7

SR209 R2 E5-2630L v3 55 8 2.9 2.6 2.5 2.4 2.3 2.2 2.1 NA NA NA 1,2,3,7

SR20A R2 E5-2603 v3 85 6 1.6 1.6 1.6 1.6 1.6 NA NA NA NA NA 1,2,3,4,5

SR1YC M1 E5-2609 v3 85 6 1.9 1.9 1.9 1.9 1.9 NA NA NA NA NA 1,2,3,4,5

SR20N R2 E5-1660 v3 140 8 3.5 3.3 3.3 3.3 3.3 3.3 3.3 NA NA NA 1,2,3,6,7

SR20B R2 E5-2608L v3 50 6 2 2 2 2 2 NA NA NA NA NA 1,2,3,7

SR20P R2 E5-1620 v3 140 4 3.6 3.6 3.6 NA NA NA NA NA NA NA 1,2,3,6,7

SR21P R2 E5-2608L v3 52 6 2 2 2 2 2 NA NA NA NA NA 1,2,3,7

SR22P C1 E5-4660 v3 120 14 2.9 2.7 2.6 2.5 2.4 2.4 2.4 2.4 2.4 2.4 1,2,3,7,

Model

Number

(W)

TDP

Intel® Turbo Boost Technology Maximum Core Frequency (GHz)

Core

# Cores

Core 3Core4Core5Core6Core7Core8Core9Core10Core

1 -2

Notes

11+

Intel® Xeon® Processor E5 v3 Product Family 9 Processor Specification Update February 2016

Table 2. Intel® Xeon® Processor E5-1600, E5-2600 and E5-4600 v3 Product Families

Turbo Bins (Sheet 3 of 3)

S-Spec

Stepping

SR22J C1 E5-4650 v3 105 12 2.8 2.6 2.5 2.4 2.4 2.4 2.4 2.4 2.4 2.4 1,2,3,7,

SR22L C1 E5-4640 v3 105 12 2.6 2.4 2.3 2.2 2.1 2.1 2.1 2.1 2.1 2.1 1,2,3,7,

SR22K C1 E5-4620 v3 105 10 2.6 2.4 2.3 2.2 2.2 2.2 2.2 2.2 2.2 NA 1,2,3,7,

SR22S C1 E5-4610 v3 105 10 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 NA 1,2,3,7,

SR22M C1 E5-4669 v3 135 18 2.9 2.7 2.6 2.5 2.4 2.4 2.4 2.4 2.4 2.4 1,2,3,7,

SR22N C1 E5-4667 v3 135 16 2.9 2.7 2.6 2.5 2.4 2.3 2.3 2.3 2.3 2.3 1,2,3,7,

SR22R M1 E5-4655 v3 135 6 3.2 3 3 3 3 NA NA NA NA NA 1,2,3,7,

SR26R C1 E5-4648 v3 105 12 2.2 2 1.9 1.9 1.9 1.9 1.9 1.9 1.9 1.9 1,2,3,7,

SR22Q M1 E5-4627 v3 135 10 3.2 3 3 3 3 3 3 3 3 NA 1,2,3,4,7,

SR233 M1 E5-2438L v3 70 10 2.3 2.1 2 2 2 2 2 2 2 NA 1,2,3,7,

SR235 R2 E5-2428L v3 55 8 2.3 2.1 2 2 2 2 2 NA NA NA 1,2,3,7,

SR234 M1 E5-1428L v3 65 8 2.5 2.3 2.2 2.2 2.2 2.2 2.2 NA NA NA 1,2,3,7,

SR232 R2 E5-2418L v3 50 6 2 2 2 2 2 NA NA NA NA NA 1,2,3,7,

SR236 R2 E5-2408L v3 45 4 1.8 1.8 1.8 NA NA NA NA NA NA NA 1,2,3,7,

Notes:

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)

Core

# Cores

Core 3Core4Core5Core6Core7Core8Core9Core10Core

1 -2

Notes

11+

Table 3. Intel® Xeon® Processor E5-1600, E5-2600 and E5-4600 v3 Product Families

Intel® AVX Turbo Bins (Sheet 1 of 3)

Intel AVX

Model Number

E5-2690 v3 2.3 12 30 3.2 3 3 3 3 3 3 3 1,2,3,7

E5-2680 v3 2.1 12 30 3.1 2.9 2.8 2.8 2.8 2.8 2.8 2.8 1,2,3,7

E5-2670 v3 2 12 30 2.9 2.7 2.6 2.6 2.6 2.6 2.6 2.6 1,2,3,7

E5-2660 v3 2.2 10 25 3.1 2.9 2.9 2.9 2.9 2.9 2.9 2.9 1,2,3,7

E5-2650 v3 2 10 25 2.8 2.6 2.6 2.6 2.6 2.6 2.6 2.6 1,2,3,7

E5-2640 v3 2.2 8 20 3.4 3.2 3.1 3 2.9 2.8 2.8 NA 1,2,3,7

E5-2630 v3 2.1 8 20 3.2 3 2.9 2.8 2.7 2.6 2.6 NA 1,2,3,7

E5-2620 v3 2.1 6 15 3.2 2.9 2.8 2.7 2.6 NA NA NA 1,2,3,7

10 Intel® Xeon® Processor E5 v3 Product Family

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)

Processor Specification Update

February 2016

Notes

Table 3. Intel® Xeon® Processor E5-1600, E5-2600 and E5-4600 v3 Product Families

Intel® AVX Turbo Bins (Sheet 2 of 3)

Intel AVX

Model Number

E5-2609 v3 1.9 6 15 NA NA NA NA NA NA NA NA 1,2,3,4,5,7

E5-2603 v3 1.3 6 15 NA NA NA NA NA NA NA NA 1,2,3,4,5,7

E5-2699 v3 1.9 18 45 3.3 3.1 3 2.9 2.8 2.7 2.6 2.6 1,2,3,7

E5-2698 v3 1.9 16 40 3.3 3.1 3 2.9 2.8 2.7 2.6 2.5 1,2,3,7

E5-2697 v3 2.2 14 35 3.3 3.1 3 2.9 2.9 2.9 2.9 2.9 1,2,3,7

E5-2695 v3 1.9 14 35 3 2.8 2.7 2.6 2.6 2.6 2.6 2.6 1,2,3,7

E5-2687W v3 2.7 10 25 3.5 3.3 3.2 3.2 3.2 3.2 3.2 3.2 1,2,3,7

E5-2685 v3 2.2 12 30 3.2 3 2.9 2.8 2.8 2.8 2.8 2.8 1,2,3,4,7

E5-2683 v3 1.7 14 35 2.7 2.5 2.5 2.5 2.5 2.5 2.5 2.5 1,2,3,7

E5-2667 v3 2.7 8 20 3.5 3.3 3.3 3.3 3.3 3.3 3.3 NA 1,2,3,7

E5-2650L v3 1.5 12 30 2.3 2.1 2.1 2.1 2.1 2.1 2.1 2.1 1,2,3,7

E5-2643 v3 2.8 6 20 3.6 3.5 3.5 3.5 3.5 NA NA NA 1,2,3,7

E5-2637 v3 3.2 4 15 3.6 3.5 3.5 NA NA NA NA NA 1,2,3,7

E5-2630L v3 1.5 8 20 2.9 2.6 2.5 2.4 2.3 2.2 2.1 NA 1,2,3,7

E5-2623 v3 2.7 4 10 3.5 3.3 3.3 NA NA NA NA NA 1,2,3,7

E5-1680 v3 2.9 8 20 3.7 3.5 3.4 3.4 3.4 3.4 3.4 NA 1,2,3,7

E5-1660 v3 2.7 8 20 3.5 3.3 3.3 3.3 3.3 3.3 3.3 NA 1,2,3,7

E5-1650 v3 3.2 6 15 3.7 3.5 3.5 3.5 3.5 NA NA NA 1,2,3,7

E5-1630 v3 3.4 4 10 3.7 3.7 3.7 NA NA NA NA NA 1,2,3,7

E5-1620 v3 3.2 4 10 3.5 3.5 3.5 NA NA NA NA NA 1,2,3,7

E5-1607 v3 2.8 4 10 NA NA NA NA NA NA NA NA 1,2,3,4,5,7

E5-1603 v3 2.5 4 10 NA NA NA NA NA NA NA NA 1,2,3,4,5,7

E5-2658 v3 1.9 12 30 2.6 2.4 2.4 2.4 2.4 2.4 2.4 2.4 1,2,3,7

E5-2648L v3 1.5 12 30 2.2 2 2 2 2 2 2 2 1,2,3,7

E5-2628L v3 1.7 10 25 2.4 2.2 2.2 2.2 2.2 2.2 2.2 2.2 1,2,3,7

E5-2618L v3 1.9 8 20 3.4 3 2.9 2.8 2.7 2.6 2.5 NA 1,2,3,7

E5-2608L v3 1.7 6 15 2 2 2 2 2 NA NA NA 1,2,3,7

E5-4660 v3 1.8 14 35 2.9 2.7 2.6 2.5 2.4 2.4 2.4 2.4 1,2,3,7

E5-4650 v3 1.8 12 30 2.8 2.6 2.5 2.4 2.4 2.4 2.4 2.4 1,2,3,7

E5-4640 v3 1.6 12 30 2.6 2.4 2.3 2.2 2.1 2.1 2.1 2.1 1,2,3,7

E5-4620 v3 1.7 10 25 2.4 2.2 2.2 2.2 2.2 2.2 2.2 2.2 1,2,3,7

E5-4610 v3 1.7 10 25 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1,2,3,7

E5-4669 v3 1.8 18 45 2.9 2.7 2.6 2.5 2.4 2.4 2.4 2.4 1,2,3,7

E5-4667 v3 1.7 16 40 2.9 2.7 2.6 2.5 2.4 2.3 2.3 2.3 1,2,3,7

E5-4655 v3 2.6 6 30 3.2 3 3 3 3 NA NA NA 1,2,3,7

E5-4648 v3 1.4 12 30 2.2 2 1.9 1.9 1.9 1.9 1.9 1.9 1,2,3,7

E5-4627 v3 2.3 10 25 3.2 3 3 3 3 3 3 3 1,2,3,4,7

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

Intel® Xeon® Processor E5 v3 Product Family 11 Processor Specification Update February 2016

Table 3. Intel® Xeon® Processor E5-1600, E5-2600 and E5-4600 v3 Product Families

Intel® AVX Turbo Bins (Sheet 3 of 3)

Intel AVX

Model Number

E5-2438L v3 1.5 10 25 2.2 2 1/9 1.9 1.9 1.9 1.9 1.9 1,2,3,7

E5-2428L v3 1.4 8 20 2.3 2.1 2 2 2 2 2 NA 1,2,3,7

E5-1428L v3 1.7 8 20 2.4 2.2 2.1 2.1 2.1 2.1 2.1 NA 1,2,3,7

E5-2418L v3 1.5 6 15 2 2 2 2 2 NA NA NA 1,2,3,7

E5-2408L v3 1.4 4 10 1.8 1.8 1.8 NA NA NA NA NA 1,2,3,7

Notes:

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

12 Intel® 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

HSE1 XNo FixVM Exit May Set IA32_EFER.NXE When IA32_MISC_ENABLE Bit 34 is Set to 1

HSE2 XNo FixIntel® QuickPath Interconnect (Intel® QPI) Layer May Report Spurious Correctable Errors

HSE3 XNo FixNTB May Incorrectly Set MSI or MSI-X Interrupt Pending Bits

HSE4 XNo FixProcessor May Issue Unexpected NAK DLLP Upon PCIe* L1 Exit

HSE5 XNo FixPECI DDR DIMM Digital Thermal Reading Returns Incorrect Value

HSE6 XNo FixIIO CSR Lnkcon2 Field Selectable_De_Emphasis Cannot Be Set For DMI2 Mode

HSE7 XNo FixPCIe* Receiver May Not Meet the Specification for AC Common Mode Voltage And Jitter

HSE8 XNo FixReceiver Termination Impedance On PCIe* 3.0 Does Not Comply With The Specification

HSE9 XNo FixPlatform Recovery After a Machine Check May Fail

HSE10 XNo FixUSB3 xHCI Not Compatible With MSIs

HSE11 XNo FixA Memory Channel With More Than 4 Ranks May Lead to a System Hang

HSE12 XNo FixWriting R3QPI Performance Monitor Registers May Fail

HSE13 XNo FixEnabling PPD May Cause Unpredictable System Behavior

HSE14 XNo Fix

HSE15 XNo FixIntel QPI Link Re-training After a Warm Reset or L1 Exit May be Unsuccessful

HSE16 XNo FixVCCIN VR Phase Shedding is Disabled

HSE17 XNo FixQuad-rank DDR4 LRDIMMs May Not be Properly Calibrated

HSE18 XNo FixPossible Non-Optimal Electrical Margins on The DDR Command Bus

HSE19 XNo FixPECI Commands During Reset May Result in Persistent Timeout Response

HSE20 XNo FixSystem May Hang When Using the TPH Prefetch Hint

HSE21 XNo FixTS1s Do Not Convey The Correct Transmitter Equalization Values During Recovery.RcvrLock

HSE22 XNo FixMSR_TEMPERATURE_TARGET MSR May Read as '0'

HSE23 XNo FixPECI RdIAMSR() Command May Fail After Core C6 State is Entered

HSE24 XNo FixProcessor May Not Enter PC3 or PC6 State

HSE25 XNo FixDisabling Intel® QPI L1 State May Cause a Machine Check On PC3 or PC6 Exit

HSE26 XNo FixCLTT May Cause BIOS To Hang On a Subsequent Warm Reset

HSE27 XNo FixSystems May Experience Uncorrectable Errors When Using 2133 MHz LRDIMM

HSE28 XNo FixPCIe* Extended Tag Field May be Improperly Set

HSE29 XNo Fix

HSE30 XNo FixThe System May Hang When a C/A Parity Error is Detected

HSE31 XNo Fix

HSE32 XNo FixEnabling Isochronous Transfers May Result in Unpredictable System Behavior

HSE33 XNo FixEnabling Secondary To Primary Snoop Overrides On NTB May Cause a Hang

HSE34 XNo FixMemory Controller tsod_present Settings Being Improperly Cleared

HSE35 XNo FixDDR4 Protocol May be Violated During C/A Parity Error Handling

HSE36 XNo FixDDR4 Power Down Timing Violation

C-1,

M-1, R-2

Status Errata

CPUID Extended Topology Enumeration Leaf May Indicate an Incorrect Number of Logical Processors

A MOV to CR3 When EPT is Enabled May Lead to an Unexpected Page Fault or an Incorrect Page Translation

A C/A Parity Error When DDR4 is Operating at 2133 MHz May Result in Unpredictable System Behavior

14 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

Table 4. Errata (Sheet 2 of 4)

Stepping

Number

HSE37 XNo FixCorrectable Memory ECC Errors May Occur at Boot

HSE38 XNo FixRemote P2P MCTP Transactions Cause Requests Timeouts

HSE39 XNo Fix

HSE40 XNo FixBT Timeouts May Cause Spurious Machine Checks

HSE41 XNo FixPCIe* Type 1 VDMs May be Silently Dropped

HSE42 XNo FixFull Duplex NTB Traffic Can Cause a System Hang

HSE43 XNo FixCONFIG_TDP_NOMINAL CSR Implemented at Incorrect Offset

HSE44 XNo FixSoftware Using Intel® TSX May Result in Unpredictable System Behavior

HSE45 XNo Fix

HSE46 XNo FixSpurious Corrected Errors May be Reported

HSE47 XNo FixPCIe* LBMS Bit Incorrectly Set

HSE48 XNo FixPower Consumed During Package C6 May Exceed Specification

HSE49 XNo FixPlatform Performance Degradation When C1E is Enabled

HSE50 XNo FixMemory Power Down Entry May Lead to Unpredictable System Behavior

HSE51 XNo FixPCIe* Correctable Error Status Register May Not Log Receiver Error at 8.0 GT/s

HSE52 XNo FixSending PECI Messages Concurrently Over Two Interfaces May Result in a System Hang

HSE53 XNo FixPlatform Recovery After a Machine Check May Fail

HSE54 XNo FixIntel® Turbo Boost Technology Does Not Behave As Expected

HSE55 XNo FixPCIe* Hot Plug Slot Status Register May Not Indicate Command Completed

HSE56 XNo FixLocal PCIe* P2P Traffic on x4 Ports May Cause a System Hang

HSE57 XNo FixNTB Operating In NTB/RP Mode May Complete Transactions With Incorrect ReqID

HSE58 XNo FixIncorrect Single-Core Turbo Ratio Limit May be Applied When Using AVX Instructions

HSE59 XNo FixLLC Error Conditions May be Dropped or Incorrectly Signaled

HSE60 XNo FixUnexpected System Behavior May Occur Following Virtualization of Some APIC Writes

HSE61 XNo FixA DDR4 C/A Parity Error in Lockstep Mode May Result in a Spurious Uncorrectable Error

HSE62 XNo FixSome IMC and Intel QPI Functions Have Incorrect PCI CAPPTR Values

HSE63 XNo FixPCIe* TLP Translation Request Errors Are Not Properly Logged For Invalid Memory Writes

HSE64 XNo FixThe TSC May be Inaccurate When Per Core P-States Are Enabled

HSE65 XNo Fix

HSE66 XNo FixPCIe* Slave Loopback May Transmit Incorrect Sync Headers

HSE67

HSE68 XNo FixData Poisoning May Not Behave as Expected

HSE69 XNo FixEnabling TRR With DDR4 LRDIMMs May Lead to Unpredictable System Behavior

HSE70 XNo FixWarm Reset May Cause PCIe Hot-Plug Sequencing Failure

HSE71 XNo FixPerformance Monitor Instructions Retired Event May Not Count Consistently

HSE72 XNo FixC/A Parity Error Injection May Cause the System to Hang

HSE73 XNo FixExcessive Fan Speed

C-1,

M-1, R-2

XNo FixConsecutive PECI RdIAMSR Commands When Core C6 is Enabled May Cause a System Hang

Status Errata

Reserving Resources For Isochronous Transfers With Non-Posted Prefetching Enabled May Cause a Hang

A Machine-Check Exception Due to Instruction Fetch May Be Delivered Before an Instruction Breakpoint

Dual HA Processors With CLTT Enabled And no Memory on Channels 0 And 1 May Hang During Warm Reset

Intel® Xeon® Processor E5 v3 Product Family 15 Processor Specification Update February 2016

Table 4. Errata (Sheet 3 of 4)

Stepping

Number

HSE74 XNo FixDisabling PCIe* Ports Prevents Package C6 Entry

HSE75 XNo FixThe System May Shut Down Unexpectedly During a Warm Reset.

HSE76 XNo Fix

HSE77 XNo FixPatrol Scrubbing of Mirrored Memory May Log Spurious Memory Errors

HSE78 XNo FixMSR_TURBO_ACTIVATION_RATIO MSR Cannot be Locked

HSE79 X No Fix Duplicate. Erratum removed

HSE80 XNo FixAn APIC Timer Interrupt During Core C6 Entry May be Lost

HSE81 XNo FixDDR4 CLK Signal May Incorrectly Float During Warm Reset or S3 State

HSE82 XNo FixDDR4 Self Ref resh Entry May Result in Memory Read Errors

HSE83 XNo Fix

HSE84 XNo FixWrites To Some Control Register Bits Ignore Byte Enable

HSE85 XNo Fix

HSE86 XNo FixWarm Reset May Cause PCIe Hot-Plug Sequencing Failure

HSE87 XNo FixPROCHOT# Assertion During Warm Reset May Cause Persistent Performance Reduction

HSE88 XNo Fix

HSE89 XNo FixPECI RdIAMSR Accesses During Boot Or Warm Reset May Cause The Processor to Hang

HSE90 XNo FixDDR4 Rank Aliasing With Heavy Memory Traffic May Lead to a System Hang

HSE91 XNo FixEarly Thermal Throttling May Occur

HSE92 XNo FixIntel QPI Link Re-training After a Warm Reset or L1 Exit May be Unsuccessful

HSE93 XNo FixPCIe* Header of a Malformed TLP is Logged Incorrectly

HSE94 XNo FixAttempting to Enter ADR May Lead to Unpredictable System Behavior

HSE95 XNo FixPECI Commands During Warm Reset May Lead to an Undefined Machine Check Error

HSE96 XNo FixSpurious Corrected Errors May be Reported

HSE97 XNo FixSystem Hang May Occur During Warm Reset Due to SMBUS Activity

HSE98 XNo FixPECI PCS Package Identifier Command Operates Incorrectly on Certain SKUs

HSE99 XNo FixIntel® Trusted Execution Technology Uses Incorrect TPM 2.0 NV Space Index Handles

HSE100 XNo FixSurprise Down Error Status is Not Set Correctly on DMI Port

HSE101 XNo FixPECI RdPkgConfig Command DRAM Services May Behave Incorrectly

HSE102 XNo Fix

HSE103 XNo Fix

HSE104 XNo FixA P-State or C-State Transition May Lead to a System Hang

HSE105 XNo FixAn IRET Instruction That Results in a Task Switch Does Not Serialize The Processor

HSE106 XNo FixLoading a Microcode Update May Result in Higher Than Expected Idle Power

C-1,

M-1, R-2

Status Errata

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

16 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

Table 4. Errata (Sheet 4 of 4)

Stepping

Number

HSE107 XNo FixA Spurious Patrol Scrub Error May be Logged

HSE108 XNo FixSome OFFCORE_RESPONSE Performance Monitoring Events May Undercount

HSE109 XNo FixEnabling ADR with UDIMMs May Result in Unpredictable System Behavior

C-1,

M-1, R-2

Status Errata

Table 5. Specification Clarifications

No. Specification clarifications

HSE1. TSX Instruction

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

1 None

Table 7. Documentation Changes

No. Documentation changes

1 Datasheet Volume 3: Statement of Volatility

Intel® Xeon® Processor E5 v3 Product Family 17 Processor Specification Update February 2016

Errata

HSE1 VM 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.

HSE2 Intel® 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.

HSE3 NTB 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.

HSE4 Processor 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.

HSE5 PECI 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.

18 Intel® 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.

HSE6 IIO 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.

HSE7 PCIe* 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.

HSE8 Receiver 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.

HSE9 Platform 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

HSE10 USB3 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.

HSE11 A 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.

HSE12 Writing 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.

HSE13 Enabling 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.

HSE14 CPUID 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.

20 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

HSE15 Intel 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.

HSE16 VCCIN 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.

HSE17 Quad-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.

HSE18 Possible 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.

HSE19 PECI 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.

HSE20 System 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.

HSE21 TS1s 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.

HSE22 MSR_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.

HSE23 PECI 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.

HSE24 Processor 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.

HSE25 Disabling 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.

22 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

HSE26 CLTT 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.

HSE27 Systems 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.

HSE28 PCIe* 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.

HSE29 A 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.

HSE30 The 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.

HSE31 A 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.

HSE32 Enabling 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.

HSE33 Enabling 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.

HSE34 Memory 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.

HSE35 DDR4 Protocol May be Violated During C/A Parity Error Handling

Problem: Due to this erratum, if the error information in the DRAM is accessed because 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.

24 Intel® 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

HSE36 DDR4 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.

HSE37 Correctable 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.

HSE38 Remote 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.

HSE39 Reserving 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.

HSE40 BT 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.

HSE41 PCIe* 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.

HSE42 Full 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.

HSE43 CONFIG_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.

HSE44 Software 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.

HSE45 A 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.

HSE46 Spurious 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.

26 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

Status: For the affected steppings, see the Summary Tables of Changes.

HSE47 PCIe* 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.

HSE48 Power 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.

HSE49 Platform 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.

HSE50 Memory 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.

HSE51 PCIe* 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.

HSE52 Sending 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.

HSE53 Platform Recovery After a Machine Check May Fail

Problem: While attempting platform recovery after a machine check (as indicated by CATERR#

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.

HSE54 Intel® 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.

HSE55 PCIe* 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.

HSE56 Local 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.

28 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

HSE57 NTB 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.

HSE58 Incorrect 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.

HSE59 LLC 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.

HSE60 Unexpected 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.

HSE61 A 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.

HSE62 Some 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.

HSE63 PCIe* 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.

HSE64 The 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.

30 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

HSE65 Dual 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.

HSE66 PCIe* 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.

HSE67 Consecutive 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.

HSE68 Data 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.

HSE69 Enabling 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.

HSE70 Warm 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.

HSE71 Performance 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.

HSE72 C/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.

HSE73 Excessive 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.

HSE74 Disabling PCIe* Ports Prevents Package C6 Entry

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

32 Intel® 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.

HSE75 The 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.

HSE76 Accessing 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.

HSE77 Patrol 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.

HSE78 MSR_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.

HSE79 Duplicate. Erratum removed.

HSE80 An 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

HSE81 DDR4 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.

HSE82 DDR4 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.

HSE83 VT-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.

HSE84 Writes 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.

HSE85 Invalid Intel® QuickData Technology XOR Descriptor Source

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.

Workaround: Ensure XOR DMA descriptor source addresses targets valid DRAM memory locations.

Status: For the affected steppings, see the Summary Tables of Changes.

34 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

HSE86 Warm 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.

HSE87 PROCHOT# 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.

HSE88 Operand-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.

HSE89 PECI 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.

HSE90 DDR4 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.

HSE91 Early 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.

HSE92 Intel 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.

HSE93 PCIe* 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.

HSE94 Attempting 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.

HSE95 PECI 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.

36 Intel® Xeon® Processor E5 v3 Product Family

during the interval from warm reset request to RESET# deassertion.

Processor Specification Update

February 2016

HSE96 Spurious 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.

corrected errors may be safely ignored.

HSE97 System 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.

HSE98 PECI 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.

HSE99 Intel® Trusted Execution Technology Uses Incorrect TPM 2.0 NV

Space Index Handles

Problem: Intel® TXT (Intel Trusted Execution Technology) uses TPM (Trusted Platform Module)

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.

HSE100 Surprise 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.

HSE101 PECI 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.

HSE102 Performance 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.

HSE103 Memory 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.

HSE104 A 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.

HSE105 An 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.

HSE106 Loading 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.

38 Intel® 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.

HSE107 A 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.

HSE108 Some 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.

HSE109 Enabling 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.

40 Intel® 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.

42 Intel® 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.

44 Intel® Xeon® Processor E5 v3 Product Family

Processor Specification Update

February 2016

Intel BX80644E52603V3 User Manual

Specifications and Main Features

Frequently Asked Questions

User Manual

Intel® Xeon® Processor E5 v3 Product Family

Revision History

Preface

Affected Documents

Nomenclature

Identification Information

Component Identification via Programming Interface

Component Marking Information

Summary Tables of Changes

Codes Used in Summary Tables

Stepping

Page

Status

Errata

Specification Changes

Specification Clarifications

Documentation Changes

Mixed Processors Within DP Platforms

Mixed Processor Consistency Requirements