Alphabetical List of All Documents
A | B | C | D | E | G | I | J | K | M | O | P | R | S | T | U | V
A
- AI Software Stack is Now Customer-Driven
A white paper from Cambrian AI Research examining the growing momentum of the Poplar software stack and ecosystem
- Approved Servers
Specification of servers approved for use with IPU-Machines and IPU-POD systems.
B
- BMC User Guide
User guide that defines the functions and the interfaces provided by the BMC software on an IPU-Machine (IPU-M2000 and Bow-2000).
- Bow Pod Direct Attach Build and Test Guide
Instructions for assembling the hardware, installing the software, and then testing a single Bow-2000 direct attach system.
- Bow Pod Expansion Guide
Instructions to expand Bow Pod systems. Expansions described: Bow Pod16 direct attach to Bow Pod64, and Bow Pod64 to larger Bow Pod systems.
- Bow Pod16 Reference Design Datasheet
Technical specifications for the Bow Pod16 reference design.
- Bow Pod256 Reference Design Datasheet
Technical specifications for the Bow Pod256.
- Bow Pod64 Reference Design Datasheet
Technical specifications for the Bow Pod64 reference design.
- Bow Pod64 Reference Design: Build and Test Guide
Instructions for assembling the hardware, installing the software, and then testing a Bow Pod64 system.
- Bow-2000 IPU-Machine Datasheet
Technical specifications for the Bow-2000 IPU-Machine.
C
- C600 Datasheet
Technical specification for the C600 PCIe Gen4 card.
- C600 PCIe Accelerator: Power and Thermal Control
Power consumption, fan control, and thermal protection mechanisms for C600 cards
- C600 PCIe SMBus Interface
SMBus specification for C600 cards
- C600 Power and Thermal Control <https://docs.graphcore.ai/projects/c600-power-thermal-control/>`__
Power consumption, fan control, and thermal protection for C600 cards
- C600 数据手册
C600 Gen4卡技术参数
- Command Line Tools
User guide for the command line tools that provide information on the current status of the connected hardware. These tools are included with the Poplar SDK.
- Contribution License Agreement (CLA)
Terms for contributing to Graphcore open source projects
- Creating Custom Operations for the IPU
An overview of the steps for implementing a custom op in each of the frameworks available in the Poplar SDK
D
- Dictionary of Graphcore Terminology
A dictionary of specialised terms related to Graphcore technology
E
G
- GCL User Guide and API Reference
Description of the classes and functions in the GCL library
- Gcore Cloud: Getting Started with IPUs
How to access IPUs and run ML applications on Gcore Cloud
- Getting Started with Bow Pod Systems
Installing the Poplar SDK and setting up the Bow Pod ready to run your application
- Getting Started with Graphcloud
How to access IPUs and run ML applications on Graphcloud
- Getting Started with IPU-POD Systems
Installing the Poplar SDK and setting up the IPU-POD ready to run your application
- Getting Started with IPU-POD4 DA and IPU-POD16 DA
Installing the Poplar SDK and setting up the IPU-POD DA ready to run your application
- Glossary
A dictionary of specialised terms related to Graphcore technology
- Graphcore Approved Servers
Specification of servers approved for use with IPU-Machines and IPU-POD systems.
- Graphcore Command Line Tools
User guide for the command line tools that provide information on the current status of the connected hardware. These tools are included with the Poplar SDK.
- Graphcore IPU Based Systems with Weka Data Platform
An example reference architecture has been developed in partnership with Weka using the Weka data platform for AI with a Graphcore Pod.
- Graphcore IPU Info Library (gcipuinfo)
User guide for the Graphcore IPU Info library (gcipuinfo). This library provides an API for monitoring and gathering information about the IPUs available in a system, and the applications using them.
- Graphcore OpenStack Reference Design for IPU-POD Systems
A reference configuration of an IPU‑POD64 deployed with OpenStack management software
- Graphcore Pod with DDN Storage
An example reference architecture has been developed in partnership with DDN storage solutions using a DDN AI400X2 storage system with a Graphcore Pod.
- Graphcore’s AI Software Stack is Now Customer-Driven
A white paper from Cambrian AI Research examining the growing momentum of the Poplar software stack and ecosystem
I
- ICU Recovery Mode (911)
Information for recovering from ICU error states, for example after a failed firmware update.
- IPU Based Systems with Weka Data Platform
An example reference architecture has been developed in partnership with Weka using the Weka data platform for AI with a Graphcore Pod.
- IPU Inference Toolkit Quick Start
Deploy a trained model to a C600 IPU
- IPU Inference Toolkit User Guide
Deploy trained models to Graphcore IPU products conveniently and quickly
- IPU Info Library (gcipuinfo)
User guide for the Graphcore IPU Info library (gcipuinfo). This library provides an API for monitoring and gathering information about the IPUs available in a system, and the applications using them.
- IPU Programmer’s Guide
An introduction to the IPU architecture, programming model and tools available
- IPU TensorFlow Serving 1 User Guide
Information on exporting models from TensorFlow 1 and running them on IPUs using TensorFlow Serving
- IPU TensorFlow Serving 2 User Guide
Information on exporting models from TensorFlow 2 and running them on IPUs using TensorFlow Serving
- IPU 推理工具包快速入门
将训练后的模型部署到 C600 IPU
- IPU 推理工具包用户指南
方便快捷地将训练好的模型部署到Graphcore IPU产品
- IPU-M2000 IPU-Machine Datasheet
Technical specifications for the IPU-M2000 IPU-Machine.
- IPU-POD Direct Attach Build and Test Guide
Instructions for assembling the hardware, installing the software, and then testing an IPU‑POD16 system.
- IPU-POD Expansion Guide
Instructions to expand IPU-POD systems. Expansions described: IPU‑POD4 direct attach to IPU‑POD16 direct attach, IPU‑POD16 direct attach to IPU‑POD64 and IPU‑POD64 to larger IPU-POD systems.
- IPU-POD128 Data Centre Checklist
A checklist to confirm requirements for data centres installing IPU‑POD128 systems.
- IPU-POD128 Reference Design Datasheet
Technical specifications for the IPU‑POD128.
- IPU-POD128 Reference Design: Build and Test Guide
Instructions for assembling the hardware, installing the software, and then testing an IPU‑POD128 system.
- IPU-POD16 Direct Attach Datasheet
Technical specifications for the IPU‑POD16.
- IPU-POD64 Reference Design: Build and Test Guide
Instructions for assembling the hardware, installing the software, and then testing an IPU‑POD64 system.
- IPU-POD64 Data Centre Checklist
A checklist to confirm requirements for data centres installing IPU‑POD64 systems.
- IPU-POD64 Reference Design Datasheet
Technical specifications for the IPU‑POD64.
- IPU-POD64 Reference Design: Installation and Integration Guide
Requirements and a high-level best practice guide for deploying the IPU‑POD64 into a data centre environment.
J
- Jupyter Notebook Quick Start
Run an application on IPUs from a Jupyter Notebook
K
- Keras with IPUs
The Graphcore implementation of TensorFlow 2 includes Keras support for IPUs
- Kubernetes IPU Device Plugin License
License details for use of the Kubernetes IPU Device Plugin
- Kubernetes IPU Device Plugin User Guide
Manage IPU resources
- Kubernetes IPU Device Plugin 用户手册
管理IPU资源
- Kubernetes IPU Operator License
License details for use of the Kubernetes IPU Operator
Kubernetes IPU Operator Release Notes
- Kubernetes IPU Operator User Guide
Kubernetes Operator support for IPUs.
M
- Memory and Performance Optimisation on the IPU
Optimising high-performance machine learning models running on the IPU
- MIT license
PopLibs, PopTorch, PopART and Poprithms are licensed under the MIT license
- Mixed-Precision Arithmetic for AI: A Hardware Perspective
The IPU’s hardware and software architectures support fast and efficient training and inference of deep learning models using mixed precision arithmetic
- Model Parallelism on the IPU with TensorFlow: Sharding and Pipelining
Ways of parallelising TensorFlow models on IPU hardware
- Model Runtime: User Guide
A library built on the Poplar runtime to enable loading and running models stored in the Poplar Exchange Format (PopEF) on the IPU.
- Monitoring Hardware Quick Start
Some tips on how to troubleshoot possible hardware problems
O
- OpenStack Reference Design for IPU-POD Systems
A reference configuration of an IPU‑POD64 deployed with OpenStack management software
- Optimising for the IPU: Computational Graph Recompilation and Executable Switching in TensorFlow
Strategies to minimise recompilation when running code on the IPU
- Optimising Temporary Memory Usage for Convolutions and Matmuls on the IPU
Using the “available memory proportion” option to optimise memory use or performance
P
- Paperspace: Getting Started with IPUs
How to access IPUs and run ML applications on Paperspace
- Pod Deployment Datacentre Checklist
A checklist to confirm requirements for data centres installing Pod systems.
- Pod Maintenance Guide
User guide that describes how to replace field-serviceable parts in a Pod system or in an IPU-Machine.
- Pod with DDN Storage
An example reference architecture has been developed in partnership with DDN storage solutions using a DDN AI400X2 storage system with a Graphcore Pod.
- PopART Quick Start
Run a PopART application
- PopART User Guide
The Poplar Advanced Runtime (PopART) for importing and executing models using the ONNX format
- PopDist and PopRun: User Guide
User guide for configuring and running distributed applications. The Poplar Distributed Configuration Library (PopDist) provides an API to prepare applications for distributed execution and PopRun is the command line utility to launch distributed applications on Pod systems.
- PopEF: User Guide
A file format for exporting and importing models to run on the IPU, and a library for managing those files
- Poplar and PopLibs API Reference
Description of the classes and functions in the Poplar and PopLibs libraries
- Poplar and PopLibs User Guide
Information on how to use the Poplar graph programming tools to write code for the IPU
- Poplar Quick Start
Run an application directly in the Poplar Graph Programming Framework
- Poplar SDK Overview
A description of the Poplar SDK with instructions for downloading and installing
- Poplar Triton Backend: User Guide
Information on the Poplar Triton backend: what it is used for, how to install it and how to use it
- PopRT User Guide
Tool for converting and compiling trained models for running on a C600 IPU
- PopRT 用户指南
工具:转换和编译在 C600 IPU 上运行的训练好的模型
- PopSparse Matrix Multiplication (Dynamic Pattern) on the IPU
A high-level description of the algorithmic design of the dynamic sparse matrix multiplication in the Graphcore PopSparse library
- PopVision Analysis Library (libpva) User Guide
The PopVision analysis library can be used for programmatic analysis of Poplar profiling information
PopVision Graph Analyser Release Notes
- PopVision Graph Analyser User Guide
The Graph Analyser can be downloaded from the PopVision tools web page
PopVision System Analyser Release Notes
- PopVision System Analyser User Guide
The System Analyser can be downloaded from the PopVision tools web page
- PopVision Trace Instrumentation Library
The PopVision trace instrumentation library provides functions to capture data used by the PopVision System Analyser
- PopXL User Guide and API (experimental)
User Guide and API reference for working with PopXL
- Porting TensorFlow 1 models to the IPU
A practical guide to porting TensorFlow models to the IPU using the Poplar SDK.
- Porting TensorFlow 2 Models Quick Start
A short description of how to port TensorFlow 2 models to the IPU, including code snippets to help
- Pre-Training and Fine-Tuning BERT for the IPU
BERT-Large implementation on Graphcore IPU-POD systems, using both TensorFlow and PyTorch
- Profiling Quick Start
Profile code that runs on the IPU as well as code that runs on the host server
- PyTorch for the IPU: User Guide
User guide and API reference for PyTorch on the IPU
- PyTorch Geometric for the IPU: User Guide
User guide and API reference for PyTorch Geometric on the IPU
- PyTorch Quick Start
Run a PyTorch application
R
- rack_tool(1)
User guide for the tool that allows you to perform operations related to one or more IPU-Machines in a rack.
S
- Scaling AI with Graphcore and Pure Storage
An example reference architecture, developed with Pure Storage, using FlashBlade storage with the IPU-POD
- Switched GW-Links in large scale Pod systems
Using switched GW-Links to connect IPU-Machines in large-scale switched Pod systems
T
- Targeting the IPU from TensorFlow 1
User guide and API reference for the IPU implementation of TensorFlow 1
- Targeting the IPU from TensorFlow 2
User guide and API reference for the IPU implementation of TensorFlow 2
- TensorFlow 1 Quick Start
Run a TensorFlow 1 application
- TensorFlow 1 User Guide
User guide and API reference for the IPU implementation of TensorFlow 1
- TensorFlow 2 Quick Start
Run a TensorFlow 2 application
- TensorFlow 2 User Guide
User guide and API reference for the IPU implementation of TensorFlow 2
- Tile Vertex Instruction Set Architecture for Mk2 IPUs
Tile vertex instruction set architecture (ISA) for Mk2 IPUs. This contains a subset of the instruction set used by the Worker threads.
- Tile Vertex Instruction Set Architecture for Mk2 IPUs with FP8 Support
Tile vertex instruction set architecture (ISA) for Mk2 IPUs with FP8 support. This contains a subset of the instruction set used by the Worker threads.
- Tutorials
Tutorials to help you get started using the Poplar SDK and Graphcore tools to run code on the IPU. You can also access these from the GitHub examples repository.
U
- Using IPUs from Docker
User guide explaining how to run applications in Docker on a Linux machine using one or more physical IPUs.
- Using MACVLAN with Kubernetes
Using Kubernetes with MACVLAN to provide access to the RDMA network interface.
V
- V-IPU Administrator Guide
User guide for administrators of data centre clusters based on the Graphcore® Virtual-IPU™ (V-IPU). The V-IPU management software provides a control plane for large-scale multi-tenanted deployments of IPUs.
V-IPU Slurm Workload Manager Release Notes
- V-IPU User Guide
User guide for users of data centre clusters based on the Graphcore® Virtual-IPU™ (V-IPU). The V-IPU management software provides a control plane for large-scale multi-tenanted deployments of IPUs.