NVIDIA A16 PCIe GPU 加速器产品简介说明书-USB迷|专注于互联网分享

2024年7月22日发(作者：巩文心)

NVIDIA A16 PCIe GPU Accelerator

Product Brief

PB-10518-001_v02 | March 2022

Document History

PB-10518-001_v02

Version

Date

June 29, 2021

March 17, 2022

Authors

VK, SM

Description of Change

Initial Release

•

Removed Secure Boot from Table 1

•

Updated Secure Boot and CEC description in

Table 3

•

Updated Table 4 “Board Environmental and

Reliability Specifications”

•

Updated “Root of Trust” section

•

Added Table 5 “Root of Trust Feature Set”

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | ii

Table of Contents

Overview ..................................................................................................................... 1

Specifications ............................................................................................................. 3

Product Specifications ................................................................................................................. 3

Environmental and Reliability Specifications ............................................................................. 5

Airflow Direction Support ........................................................................................... 6

Product Features ....................................................................................................... 7

PCI Express Interface Specifications .......................................................................................... 7

PCIe Support ............................................................................................................................. 7

Polarity Inversion and Lane Reversal Support ....................................................................... 7

Root of Trust ................................................................................................................................. 7

nvidia-smi ................................................................................................................................. 8

SMBPBI ..................................................................................................................................... 8

Power Connector Placement ....................................................................................................... 9

CPU 8-Pin to PCIe 8-Pin Power Adapter .............................................................................. 10

Support Information ................................................................................................. 11

Certifications .............................................................................................................................. 11

Agencies ...................................................................................................................................... 11

Languages .................................................................................................................................. 12

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | iii

List of Figures

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Single Board with Four GPUs ...................................................................................... 1

NVIDIA A16 PCIe Card .................................................................................................. 2

A16 Airflow Direction .................................................................................................... 6

CPU 8-Pin Power Connector ....................................................................................... 9

CPU 8-Pin to PCIe 8-Pin Power Adapter .................................................................. 10

List of Tables

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Product Specifications ................................................................................................. 3

Memory Specifications ................................................................................................. 4

Software Specifications ................................................................................................ 4

Board Environmental and Reliability Specifications .................................................. 5

Root of Trust Feature Set ............................................................................................. 8

SMBPBI Commands ..................................................................................................... 9

Supported Auxiliary ...................................................................................................... 9

Languages Supported ................................................................................................ 12

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | iv

Overview

NVIDIA A16 is a PCI Express Gen4 graphics processing unit (GPU) card that is ideal for

providing high-user density for Virtual Desktop Infrastructure (VDI) environments. It is a full

height, full length (FHFL) design with four GPUs on a single board. The A16 is a dual-slot card

featuring 64 GB of GDDR6 memory and a 250 W maximum power limit. The A16 also supports

x16 PCIe Gen4 connectivity. It is a passively cooled card with a superior thermal design that

requires system airflow to operate and handles challenging ambient environments with ease

(NEBS-3 capable).

throughput and frame buffer for the best user experience in a VDI environment using NVIDIA

Virtual PC (vPC) software. Video transcoding and Android

™

cloud gaming are among the other

workloads that can take advantage of the multiple encoders and decoders on the A16 GPU.

The quad GPU design enables the highest frame buffer, encoder, and decoder density in a

dual-slot form factor for VDI use cases.

Refer to the following website for the latest list of servers qualified with NVIDIA A16:

/en-us/data-center/tesla/tesla-qualified-servers-catalog/

Figure 1. Single Board with Four GPUs

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 1

Figure 2. NVIDIA A16 PCIe Card

NVIDIA A16 PCIe GPU Accelerator

Overview

PB-10518-001_v02 | 2

Specifications

Product Specifications

Table 1 through Table 3 provide the product, memory, and software specifications for the

NVIDIA A16 PCIe card.

Table 1.

Specification

Product SKU

Product Specifications

NVIDIA A16

PG171 SKU 200

NVPN: 699-2G171-0200-xxx

250 W default

250 W maximum

235 W MaxQ (maximum performance per watt)

195 W minimum

Total board power

Thermal solution

Mechanical Form Factor

GPU SKU

Board PCI Device IDs

Passive

Full-height, full-length (FHFL) 10.5”, dual-slot

GA107-890 (quantity four)

Device ID: 0x25B6

Vendor ID: 0x10DE

Sub-Vendor ID: 0x10DE

Sub-System ID: 0x14A9

Switch PCI Device IDs Device ID: 0x4123

Vendor ID: 0x15B3

Sub-Vendor ID: 0x15B3

Sub-System ID: 0x1103

GPU clocks

VBIOS

PCI Express interface

Base: 1312 MHz

Boost: 1755 MHz

EEPROM size: 8 Mbit

UEFI: Supported

Physical x16 PCIe lanes

PCIe Gen4 x16 or Gen3 x16 supported

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 3

Specifications

Specification

Zero Power

Weight

NVIDIA A16

Lane and polarity reversal supported

Not supported

Board: 1088 Grams (excluding bracket and extenders)

Bracket with screws: 20 Grams

Long offset extender: 64 Grams

Straight extender: 39 Grams

Table 2.

Specification

Memory clock

Memory type

Memory size

Memory Specifications

Description

6.25 GHz

GDDR6

64 GB (16 GB per GPU)

128 bits

Up to 4x 200 GB/sec

Memory bus width

Peak memory bandwidth

Table 3.

Specification

SR-IOV support

Software Specifications

Description

Supported: 16 VF (virtual functions) per GPU

BAR0: 16 MiB

BAR1: 16 GiB

BAR3: 32 MiB

BAR address (physical function)

BAR address (virtual function)

BAR0: 4 MiB, (16 VF x 256 KiB per VF)

BAR1: 32 GiB, 64-bit (16 VF x 2 GiB per VF)

BAR3: 512 MiB, 64-bit (16 VF x 32 MiB per VF)

Message signaled interrupts

ARI Forwarding

Driver support

Secure Boot

CEC Firmware

NVIDIA

CUDA

support

Virtual GPU software support

NVIDIA

NGC-Ready

™

test suite

PCI class code

MSI-X: Supported

MSI: Not supported

Supported

R470 or later

Supported (see “Root of Trust” section)

v4.01 or later (for CEC-enabled cards)

CUDA 11.4 or later

Supports vGPU 13.0 or later

Supports NVIDIA vPC, vApps, NVIDIA RTX

™

vWS, vCS

NGC-Next Certification 2.x or later

0x03 – Display Controller

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 4

Specifications

Specification

PCI sub-class code

ECC support

SMBus (8-bit address)

Description

0x02 – 3D Controller

Enabled (by default). Can be disabled via software

GPU1: 0x9E (write), 0x9F (read)

GPU2: 0x9C (write), 0x9D (read)

GPU3: 0x9A (write), 0x9B (read)

GPU4: 0x98 (write), 0x99 (read)

Reserved I2C addresses

SMBus direct access

SMBPBI SMBus Post-Box Interface)

Note:

CEC1: 0xAA (write)

CEC2: 0xAC (write)

Supported

The KiB, MiB and GiB notation emphasizes the “power of two” nature of the values. Thus,

•

256 KiB = 256 x 1024

•

16 MiB = 16 x 1024

•

64 GiB = 64 x 1024

See Section “Root of Trust” of this product brief

Environmental and Reliability Specifications

Table 4 provides the environment conditions specifications for the A16 PCIe card.

Table 4.

Specification

Board Environmental and Reliability Specifications

Description

0 °C to 50 °C

-5 °C to 55 °C

-40 °C to 75 °C

5% to 93% relative humidity

5% to 85% relative humidity

5% to 95% relative humidity

Uncontrolled environment:

677,793 hours at 35 °C

Controlled environment:

874,410 hours at 35 °C

Ambient operating temperature

Ambient operating temperature (short term)

Storage temperature

Operating humidity (short term)

Operating humidity

Storage humidity

Mean time between failures (MTBF)

Notes: Specifications in this table are applicable up to 6000 feet.

A period not more than 96 hours consecutive, not to exceed 15 days per year.

Some environmental stress with limited maintenance (GF35).

No environmental stress with optimum operation and maintenance (GB35).

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 5

Airflow Direction Support

The NVIDIA A16 PCIe card employs a bidirectional heat sink, which accepts airflow either left-

to-right or right-to-left directions.

Figure 3. A16 Airflow Direction

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 6

Product Features

PCI Express Interface Specifications

The following subsections describe the PCIe interface specifications for the A16 PCIe card.

PCIe Support

The A16 card supports PCIe Gen4. Gen4 x16 interface should be used when connecting to the

A16 PCIe card. For optimal GPU performance, a Gen4 ×16 connection is recommended.

However, Gen3 ×16 link connection is supported as well.

Polarity Inversion and Lane Reversal Support

Lane Polarity Inversion, as defined in the PCIe specification, is supported on the A16 PCIe

card.

Lane Reversal, as defined in the PCIe specification, is supported on the A16 PCIe card. When

reversing the order of the PCIe lanes, the order of both the Rx lanes and the Tx lanes must be

reversed.

Root of Trust

The NVIDIA A16 GPU has a primary root of trust within the GPU chip that provides the

following:

 Secure boot

 Secure firmware upgrade

 Firmware rollback protection

 Ability to disable In-band firmware update (established after each GPU reset)

 Secure application processor recovery

In addition, NVIDIA offers some A16 PCIe boards with an onboard CEC1712 chip, acting as a

secondary root of trust, extending the security capabilities allowing for firmware attestation,

key revocation, and out-of-band firmware updates. The CEC1712 device authenticates the

contents of the GPU firmware ROM before permitting the GPU to boot from its ROM. For

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 7

Product Features

CEC1712-enabled cards, the root of trust feature occupies up to two I2C addresses (in addition

to the SMBus addresses). I2C addresses 0xAA and 0xAC should therefore be avoided for

system use.

Identification of the two variants of A16 PCIe boards (with or without CEC1712) can be done

using the 900-level part number on the back of the GPU or running the nvidia-smi -q

command.

 900-2G171-XXXX-1XX A16 GPUs without CEC1712 (secondary root of trust)

 900-2G171-XXXX-0XX A16 GPUs with CEC1712 (secondary root of trust)

The following table shows the features that are available using the primary and secondary root

of trust.

Table 5.

Features

Secure Boot

Root of Trust Feature Set

Primary Root of Trust

within GPU Chip

Yes

Secondary Root of Trust Using

External CEC Chip on Board

Yes

Secure Firmware Upgrade

Firmware Rollback Protection

In-Band Firmware Update Disable

Key Revocation

Firmware Attestation

Notes:

“In-Band Firmware Update Disable” feature must be established after every GPU reset.

nvidia-smi

nvidia-smi is an in-band monitoring tool provided with the NVIDIA driver and can be used to

set the maximum power consumption with driver running in persistence mode. An example

command to enable Max-Q with a power limit of 195 W is shown:

nvidia-smi -pm 1

nvidia-smi -pl 195

To restore the A16 back to its default TDP power consumption, either the driver module can be

unloaded and reloaded, or the following command can be issued:

nvidia-smi -pl 250

SMBPBI

An out-of-band channel exists through the SMBus Post-Box Interface (SMBPBI) protocol to

set the power limit of the GPU. This also requires that the NVIDIA driver be loaded for full

functionality. Max-Q mode can be enabled through the following asynchronous command:

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 8

Product Features

Table 6.

Specification

Opcode

Arg1

Arg2

SMBPBI Commands

Value

10h – Submit/poll asynchronous request

0x01 – Set total GPU power limit

0x00

Power Connector Placement

The PCIe card provides a CPU 8-pin power connector on the east edge of the board.

Figure 4. CPU 8-Pin Power Connector

CPU 8-Pin Power Connector

Table 7 lists supported auxiliary power connections for the NVIDIA A16 PCIe card.

Table 7.

CPU 8-pin

Supported Auxiliary

PSU Cable

1x CPU 8-pin cable

Board Connector

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 9

Product Features

CPU 8-Pin to PCIe 8-Pin Power Adapter

Figure 5 lists the pin assignments of the power adapter. Consult NVIDIA Applications

Engineering for qualified suppliers of the power adapter.

The CPU 8-pin smart power adapter NVPN is 030-1233-000.

Figure 5. CPU 8-Pin to PCIe 8-Pin Power Adapter

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 10

Support Information

Certifications

 Windows Hardware Quality Lab (WHQL):

•

Certified Windows 10

Certified Windows Server 2008 R2, Windows Server 2012 R2

 Ergonomic requirements for office work W/VDTs (ISO 9241)

 EU Reduction of Hazardous Substances (EU RoHS)

 Joint Industry guide (J-STD) / Registration, Evaluation, Authorization, and Restriction of

Chemical Substance (EU) – (JIG / REACH)

 Halogen Free (HF)

 EU Waste Electrical and Electronic Equipment (WEEE)

Agencies

 Australian Communications and Media Authority and New Zealand Radio Spectrum

Management (RCM)

 Bureau of Standards, Metrology, and Inspection (BSMI)

 Conformité Européenne (CE)

 Federal Communications Commission (FCC)

 Industry Canada - Interference-Causing Equipment Standard (ICES)

 Korean Communications Commission (KCC)

 Underwriters Laboratories (cUL, UL)

 Voluntary Control Council for Interference (VCCI)

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 11

Support Information

Languages

Table 8.

Languages

English (US)

English (UK)

Arabic

Chinese, Simplified

Chinese, Traditional

Czech

Danish

Dutch

Finnish

French (European)

German

Greek

Hebrew

Hungarian

Italian

Japanese

Korean

Norwegian

Polish

Portuguese (Brazil)

Portuguese (European/Iberian)

Russian

Slovak

Slovenian

Spanish (European)

Spanish (Latin America)

Swedish

Thai

Turkish

Note:

Microsoft Windows 10, Windows Server 2008 R2, Windows Server 2012 R2, and Windows

Languages Supported

Windows

Yes

Linux

Yes

2016 are supported.

NVIDIA A16 PCIe GPU Accelerator PB-10518-001_v02 | 12

Notice

This document is provided for information purposes only and shall not be regarded as a warranty of a certain functionality, condition, or quality of a

product. NVIDIA Corporation (“NVIDIA”) makes no representations or warranties, expressed or implied, as to the accuracy or completeness of the

information contained in this document and assumes no responsibility for any errors contained herein. NVIDIA shall have no liability for the

consequences or use of such information or for any infringement of patents or other rights of third parties that may result from its use. This document

is not a commitment to develop, release, or deliver any Material (defined below), code, or functionality.

NVIDIA reserves the right to make corrections, modifications, enhancements, improvements, and any other changes to this document, at any time

without notice.

Customer should obtain the latest relevant information before placing orders and should verify that such information is current and complete.

NVIDIA products are sold subject to the NVIDIA standard terms and conditions of sale supplied at the time of order acknowledgement, unless otherwise

agreed in an individual sales agreement signed by authorized representatives of NVIDIA and customer (“Terms of Sale”). NVIDIA hereby expressly

objects to applying any customer general terms and conditions with regards to the purchase of the NVIDIA product referenced in this document. No

contractual obligations are formed either directly or indirectly by this document.

NVIDIA products are not designed, authorized, or warranted to be suitable for use in medical, military, aircraft, space, or life support equipment, nor

in applications where failure or malfunction of the NVIDIA product can reasonably be expected to result in personal injury, death, or property or

environmental damage. NVIDIA accepts no liability for inclusion and/or use of NVIDIA products in such equipment or applications and therefore such

inclusion and/or use is at customer’s own risk.

NVIDIA makes no representation or warranty that products based on this document will be suitable for any specified use. Testing of all parameters of

each product is not necessarily performed by NVIDIA. It is customer’s sole responsibility to evaluate and determine the applicability of any information

contained in this document, ensure the product is suitable and fit for the application planned by customer, and perform the necessary testing for the

application in order to avoid a default of the application or the product. Weaknesses in customer’s product designs may affect the quality and reliability

of the NVIDIA product and may result in additional or different conditions and/or requirements beyond those contained in this document. NVIDIA

accepts no liability related to any default, damage, costs, or problem which may be based on or attributable to: (i) the use of the NVIDIA product in any

manner that is contrary to this document or (ii) customer product designs.

No license, either expressed or implied, is granted under any NVIDIA patent right, copyright, or other NVIDIA intellectual property right under this

document. Information published by NVIDIA regarding third-party products or services does not constitute a license from NVIDIA to use such products

or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other

intellectual property rights of the third party, or a license from NVIDIA under the patents or other intellectual property rights of NVIDIA.

Reproduction of information in this document is permissible only if approved in advance by NVIDIA in writing, reproduced without alteration and in full

compliance with all applicable export laws and regulations, and accompanied by all associated conditions, limitations, and notices.

THIS DOCUMENT AND ALL NVIDIA DESIGN SPECIFICATIONS, REFERENCE BOARDS, FILES, DRAWINGS, DIAGNOSTICS, LISTS, AND OTHER

DOCUMENTS (TOGETHER AND SEPARATELY, “MATERIALS”) ARE BEING PROVIDED “AS IS.” NVIDIA MAKES NO WARRANTIES, EXPRESSED, IMPLIED,

STATUTORY, OR OTHERWISE WITH RESPECT TO THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT,

MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. TO THE EXTENT NOT PROHIBITED BY LAW, IN NO EVENT WILL NVIDIA BE LIABLE

FOR ANY DAMAGES, INCLUDING WITHOUT LIMITATION ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE, OR CONSEQUENTIAL DAMAGES,

HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, ARISING OUT OF ANY USE OF THIS DOCUMENT, EVEN IF NVIDIA HAS BEEN

ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Notwithstanding any damages that customer might incur for any reason whatsoever, NVIDIA’s

aggregate and cumulative liability towards customer for the products described herein shall be limited in accordance with the Terms of Sale for the

product.

Trademarks

NVIDIA, the NVIDIA logo, CUDA, NGC-Ready, and NVIDIA RTX are trademarks and/or registered trademarks of NVIDIA Corporation in the U.S. and

other countries. Other company and product names may be trademarks of the respective companies with which they are associated.

Google

Android, Android TV, Google Play, and the Google Play logo are trademarks of Google, Inc.

NVIDIA Corporation | 2788 San Tomas Expressway, Santa Clara, CA 95051