Blog

MOST RECENT POSTS

View all recent

Jul 14, 2025

June Papers: Gradient Norms, LLM Reasoning and Video Generation

Jun 10, 2025

May Papers: Parallel scaling, Evolving code, Understanding LLM reasoning

May 12, 2025

April Papers: Motion Prompting, Mamba Reasoning and Modeling Rewards

Apr 28, 2025

March Papers: De-Norming, Skill-Scaling, Over-Training, Drug-Generating

Mar 06, 2025

February Papers: Learning to Scale

Feb 04, 2025

January Papers: More Like “Reas-anuary Papers”

Jan 13, 2025

Spend Your FLOPs Wisely: December Papers

Dec 10, 2024

An LLM feast: November papers

Nov 13, 2024

Improving image generation and making LLMs think: October Papers

Nov 07, 2024

Graphcore begins hiring drive with 75 new jobs

Oct 18, 2024

Proper conditioning: September papers of the month

Sep 13, 2024

Hallucinations, Quantisations and Test-Time Computations: POTM August

Sep 12, 2024

All about scaling: July papers of the month

Aug 06, 2024

Graphcore joins Ultra Ethernet Consortium

Jul 11, 2024

Graphcore joins SoftBank Group to build next generation of AI compute

Jul 04, 2024

Mamba-2 & Matmul-free Models: June Papers of the Month

Jun 06, 2024

xLSTM, Schedule-Free Optimizers, Multi-token prediction: POTM May

May 15, 2024

Royal Society Fellowships for Graphcore’s Knowles and Fitzgibbon

May 02, 2024

TriForce, QuaRot, Mixture-of-Depths: Papers of the Month (Apr 2024)

Mar 07, 2024

Longer RoPEs and Better Quantisation: Papers of the Month (Feb 2024)

Feb 05, 2024

Great Teachers and Beyond Chinchilla: Papers of the Month (Jan 2024)

Dec 14, 2023

Graphcore Presenting FP8 research at NeurIPS 2023

Nov 07, 2023

IPU delivers 10X acceleration for DNA and protein sequence alignment

Nov 01, 2023

Gcore extends its IPU cloud offering to the USA

Oct 23, 2023

AI Safety Summit - First Do No Harm (Primum non nocere)

Oct 04, 2023

Deploy an API on Graphcore IPUs in less than 10 lines of code

Sep 20, 2023

Running Julia on Graphcore IPUs

Sep 14, 2023

Hadean turns to IPU to develop UK-sovereign AI capabilities for national security

Sep 12, 2023

Graphcore-Pienso partnership wins CogX Best Innovation: NLP Award

Sep 06, 2023

Graphcore joins the PyTorch Foundation

Aug 30, 2023

Making FP16 and FP8 easy to use with our new unit scaling library

Aug 11, 2023

Text summarization with BART Large on IPU

Aug 09, 2023

Fine-tune OpenAI's Whisper Automatic Speech Recognition (ASR) model

Aug 01, 2023

Llama 2: run Meta’s Open Source Large Language Model for free on IPUs

Jul 27, 2023

Fine-tuning Flan-T5 XXL - the powerful and efficient LLM

Jul 26, 2023

Graphcore expands AI tools ecosystem as UbiOps adds IPU support

Jul 19, 2023

Extending our PyTorch Geometric on IPUs support for large and heterogeneous graphs

Jul 19, 2023

Fraud detection using Graph Neural Networks with PyTorch Geometric

Jul 05, 2023

Graphium: an IPU-ready Python Library for training molecular GNNs at scale

Jul 05, 2023

Multitask Molecular Modelling with Graphium on the IPU

Jun 26, 2023

Gcore increases IPU cloud offering with new UK cluster

Jun 09, 2023

NLP Cloud adding IPU-powered models to its AI-as-a-Service platform

May 31, 2023

OpenAssistant fine-tuned Pythia-12B: open-source ChatGPT alternative

May 30, 2023

Flan-T5: sweet results with the smaller, more efficient LLM

May 30, 2023

Running Flan-T5-XL inference in Float16 for IPU - how we did it

May 26, 2023

Using generative AI commercially? Check it passes these three tests

May 23, 2023

Small language models, high performance: DeBERTa and the Future of NLU

May 19, 2023

PackedBERT: How to accelerate NLP tasks for Transformers with packing

May 16, 2023

Pienso offers efficient LLM access for business, powered by cloud IPUs

May 04, 2023

Argonne National Lab makes Bow IPUs available to researchers worldwide

Apr 26, 2023

Dolly 2.0 - open source language model with ChatGPT-like interactivity

Apr 25, 2023

How to use OpenAI’s Whisper for speech recognition

Apr 17, 2023

Accelerating Block Sparse Matrix Multiplication on IPUs with PopSparse

Apr 05, 2023

Accelerating PyG on IPUs: Unleash the Power of Graph Neural Networks

Apr 05, 2023

Getting started with PyTorch Geometric (PyG) on Graphcore IPUs

Apr 05, 2023

Graphcore users can now build and run GNNs with PyTorch Geometric

Mar 31, 2023

Open letter to the UK Prime Minister on investment in Exascale compute

Mar 29, 2023

Simple FP16 and FP8 training with unit scaling

Mar 28, 2023

How Graphcore's BERT breakthrough made Pienso's LLMs more efficient

Mar 24, 2023

Fine-tune GPT-J: a cost-effective GPT-4 alternative for many NLP tasks

Mar 24, 2023

Demo: How to fine-tune GPT-J for NLP tasks - text entailment example

Feb 16, 2023

Graphcore helps PNNL accelerate 3D molecular modelling with GNNs

Feb 09, 2023

Graphcore and Aleph Alpha partnership expands into semantic search

Jan 31, 2023

What GNNs are great at - and why Graphcore IPUs are great at GNNs

Jan 12, 2023

Paperspace and Graphcore launch ‘pay-as-you-grow' Gradient Notebooks

Jan 12, 2023

Rapidly building NLP applications with Hugging Face and IPUs

Dec 21, 2022

How to run Stable Diffusion Inference on IPUs with Paperspace

Dec 20, 2022

Poplar SDK 3.1 now available

Nov 29, 2022

Graphcore launches C600 PCIe card for AI compute

Nov 21, 2022

What's the Best AI Hardware for Graph Neural Networks?

Nov 16, 2022

Graphcore and Aleph Alpha demonstrate 80% sparsified AI model

Nov 09, 2022

Argonne National Laboratory adds Graphcore Bow IPU AI system

Oct 14, 2022

Getting started with Poplar Triton Backend

Oct 12, 2022

Preparing for the future of foundation models

Oct 04, 2022

Getting started with TensorFlow Serving for IPU

Oct 04, 2022

Training large models more stably with automatic loss scaling

Sep 27, 2022

Compute giant HPE extends Graphcore’s global reach

Sep 06, 2022

Graphcore open-sources Argon Streams AV1 video decoder verification tool

Aug 31, 2022

Do neural networks dream of falling snow?

Aug 10, 2022

Paperspace and Graphcore partner to deliver free IPUs to developers

Aug 10, 2022

Deep dive: vision transformers on Hugging Face Optimum Graphcore

Aug 10, 2022

Getting started with IPUs on Paperspace

Aug 09, 2022

Pienso and Graphcore empower business with deeper, faster AI insights

Jul 27, 2022

Improving journey time predictions with GNNs on the Graphcore IPU

Jul 21, 2022

Man Group unlocks massively parallel option pricing with Graphcore IPU

Jul 18, 2022

Graphcore's Groupbert research on show at ICML 2022 workshop

Jul 06, 2022

ETRI, Graphcore partner on high-efficiency software for large models

Jul 05, 2022

Graphcore and AMD propose FP8 AI standard with Qualcomm support

Jun 29, 2022

Baidu results underscore big MLPerf gains for Graphcore

Jun 27, 2022

Graphcore IPU powers brain-inspired deep learning at Kindai University

Jun 23, 2022

Computers for Superhuman Cognition: Simon Knowles at AICAS 2022

Jun 22, 2022

G-Core Labs IPU Cloud service now available

Jun 14, 2022

Accelerating and scaling temporal graph networks on the Graphcore IPU

Jun 08, 2022

Graphcore and Aleph Alpha partner on large, multi-modal AI models

May 30, 2022

EPCC - UK’s leading supercomputing centre - adds Graphcore IPU System

May 30, 2022

Graphcore to accelerate fusion energy research at Hartree Centre

May 30, 2022

Graphcore IPUs adopted in Argonne National Lab’s AI Testbed

May 26, 2022

Graphcore and Hugging Face launch new lineup of IPU-ready transformers

May 20, 2022

Announcing Graphcore’s Accelerated Computing Academy

May 19, 2022

Graphcore and Tencent Cloud launch IPU Preview service

May 13, 2022

Lenovo servers now available for Graphcore IPU systems

May 12, 2022

Graphcore translates to price and performance gains for Twigfarm

Apr 27, 2022

Computing for Ultra-Intelligence: Simon Knowles at WAICF

Apr 26, 2022

LabGenius speeds up AI-based drug discovery with Graphcore IPUs

Apr 26, 2022

Accelerating Conversational AI with Gridspace and Graphcore

Apr 06, 2022

Optimising IPU Applications Using the PopVision Analysis Tools

Mar 15, 2022

Germany’s National HPC Centre at KIT installs Graphcore IPU system

Mar 09, 2022

AI for Simulation: how Graphcore is helping transform traditional HPC

Mar 03, 2022

The WoW Factor: Graphcore systems get huge power and efficiency boost

Mar 03, 2022

Graphcore announces roadmap to Ultra Intelligence AI Supercomputer

Mar 03, 2022

PNNL and Graphcore partner to advance AI for scientific research

Feb 15, 2022

Graphcore and NEC partner to deliver advanced AI compute

Feb 02, 2022

G-Core Labs launches European AI cloud powered by Graphcore IPU

Jan 25, 2022

Building a better BERT: achieving 2x training speed with GroupBERT

Jan 17, 2022

Accelerating ResNet-50 Training on the IPU for MLPerf

Jan 12, 2022

Pure Storage and Graphcore release first reference architecture

Dec 20, 2021

Sensoro chooses Graphcore IPU AI Technology for safer, greener towns and cities

Dec 20, 2021

Running End-to-End YOLO Object Detection on the IPU

Dec 16, 2021

Accelerating Computer Vision: scaling EfficientNet to large IPU-Pod Systems

Dec 13, 2021

Baidu PaddlePaddle now fully supported by Graphcore IPU

Dec 12, 2021

Get Started with Baidu PaddlePaddle on IPU Hardware

Dec 09, 2021

Microscopic World: Graphcore IPU powers DP Technology’s molecular dynamics simulation platform

Dec 09, 2021

Leveraging the IPU for Molecular Dynamics Simulation

Dec 06, 2021

Building Large Models on IPU

Dec 03, 2021

Exploiting Tile-Level Parallelism on Graphcore IPUs

Dec 01, 2021

Performance at Scale: Graphcore’s Latest MLPerf Training Results

Nov 30, 2021

Getting Started with Hugging Face Transformers for IPUs with Optimum

Nov 24, 2021

Using the IPU to accelerate Spatial Interpolation Algorithm of Weather Data

Nov 23, 2021

Environmental AI: how the IPU is helping Agilor watch our water

Nov 16, 2021

Climate Change: Foreseeing the Unexpected with Graphcore IPUs

Nov 09, 2021

Graphcore Poplar SDK Container Images Now Available on Docker Hub

Oct 28, 2021

Differential Privacy and AI: Better Protecting Health Information with IPUs

Oct 27, 2021

Graphcore advances Scientific Research in New US Supercomputing Project

Oct 21, 2021

The Next Big Thing: Introducing IPU-POD128 and IPU-POD256

Oct 19, 2021

VMware and Graphcore partner to bring AI resource disaggregation to enterprise environments

Oct 18, 2021

Graphcore advances in Southeast Asia with new Singapore base

Sep 30, 2021

Graphcore adds leading storage providers to Elite Partner Program

Sep 14, 2021

Hugging Face and Graphcore partner for IPU-optimized Transformers

Aug 25, 2021

Designing the Colossus Mk2 IPU: Simon Knowles at Hot Chips 2021

Aug 11, 2021

Insuring success: Tractable accelerates accident and disaster recovery with Graphcore IPU

Aug 10, 2021

Graphcore opens Gdańsk office to strengthen R+D and customer support

Jul 27, 2021

Introducing the Initial Release of PyTorch Lightning for Graphcore IPUs

Jul 27, 2021

Getting Started with PyTorch Lightning for the IPU

Jul 22, 2021

Atos and Graphcore: bringing AI compute to the world

Jul 12, 2021

Introducing Packed BERT for 2x Training Speed-up in Natural Language Processing

Jul 09, 2021

Université de Paris accelerates Cosmology applications with Graphcore IPUs

Jul 08, 2021

Training Sparse, Large-Scale Language Models on Graphcore’s IPU

Jul 02, 2021

Graphcore IPUs to power Korea Telecom AI Data Centre

Jun 30, 2021

Raising the bar: Graphcore's first MLPerf results

Jun 28, 2021

Removing Batch Dependence in CNNs by proxy-normalising Activations

Jun 28, 2021

How we made EfficientNet more efficient

Jun 08, 2021

Graphcore turbocharges multi-horizon financial forecasting for Oxford-Man Institute

Jun 07, 2021

NHN chooses Graphcore for Korean AI Services

Jun 03, 2021

Colossus, Poplar and Tommy Flowers...

Jun 01, 2021

How to change the world with AI: Simon Knowles speaks on Pieter Abbeel Podcast

May 31, 2021

Graphcore talks Scaling up AI on Weights and Biases Podcast

May 11, 2021

University of Bristol solves scientific supercomputing problems with IPU systems

May 06, 2021

BERT-Large Training on the IPU explained

May 05, 2021

Accelerating America with new Graphcore partners

Apr 27, 2021

Graphcore Japan hits the ground running with SCSK and HPC Systems

Apr 22, 2021

Graphcore dives deep into software architecture layers on Practical AI Podcast

Apr 22, 2021

Graphcore advances sparse compute in new EU research project

Apr 21, 2021

Graphcore poised for explosive growth in China with partner Digital China

Mar 25, 2021

Supermicro Ultra servers now available in Graphcore IPU-POD

Mar 24, 2021

Start your innovation journey with IPUs: Introducing the Graphcore IPU-POD16

Mar 17, 2021

Optimising AI Performance with Graphcore PopVision Analysis Tools

Jan 26, 2021

Introducing Graphcloud: Graphcore’s MK2 IPU-POD AI cloud service with Cirrascale

Jan 26, 2021

Graphcore Academic Programme: supporting and accelerating AI innovation

Jan 06, 2021

Machine Intelligence Research Directions in 2021

Dec 29, 2020

Graphcore raises $222 million in Series E Funding Round

Dec 14, 2020

University of Massachusetts Accelerates COVID-19 Modelling on the IPU

Dec 09, 2020

Graphcore sets new AI Performance Standards with MK2 IPU Systems

Dec 09, 2020

Introducing the Production Release of PyTorch for the IPU

Nov 26, 2020

Growing Graphcore in Germany

Nov 23, 2020

Reinventing Accelerated Computing for AI: Graphcore at SC20

Nov 12, 2020

Training Neural Networks in Low-Dimensional Random Bases

Sep 29, 2020

Graphcore announces support for ODLA

Sep 25, 2020

The Future of AI Chips lies in Software

Sep 22, 2020

Announcing Graphcore’s Global Partner Program

Aug 28, 2020

University of Bristol tackles Challenges in Particle Physics with Graphcore’s IPU

Jul 23, 2020

Accelerating Text-To-Speech Models with the IPU

Jul 15, 2020

Introducing 2nd Generation IPU Systems for AI at Scale

Jul 08, 2020

Graphcore China launches IPU Developer Cloud

Jul 07, 2020

Delving Deep into Modern Computer Vision Models

Load More