ALCF AI Testbed

The ALCF AI Testbed provides an infrastructure for the next-generation of AI-accelerator machines. 

The AI Testbed aims to help evaluate the usability and performance of machine learning-based high-performance computing applications running on these accelerators. The goal is to better understand how to integrate with existing and upcoming supercomputers at the facility to accelerate science insights.

We are currently offering allocations on our Groq, Graphcore Bow IPUs, Cerebras CS-2, and SambaNova DataScale systems. 

AI-Testbed Links

Request an Allocation on Groq, Graphcore, Cerebras and/or SambaNova

Apply for Time

AI Testbed Guide

Support

AI Testbed Training

Training

Email us for more information

Learn More

Systems

GroqRack (Available for Allocation Requests)

GroqRack Inference

  • System Size: 72 Accelerators (9 nodes  x 8 Accelerators per node) 
  • Compute Units per Accelerator: 5120 vector ALUs
  • Performance of a single accelerator (TFlops): >188 (FP16) >750 (INT8)
  • Software Stack Support:  GroqWare SDK, ONNX
  • Interconnect: RealScale TM
GroqRack at ALCF

Cerebras CS-2 (Available for Allocation Requests)

Cerebras CS-2 Wafer-Scale Cluster WSE-2

  • System Size: 2 Nodes (each with a Wafer scale engine)  including Memory-X and Swarm-X
  • Compute Units per Accelerator: 850,000 Cores
  • Performance of a single accelerator (TFlops): >5780 (FP16) 
  • Software Stack Support: Cerebras SDK, Tensorflow, Pytorch
  • Interconnect: Ethernet-based
Cerebras Wafer-Scale Cluster

SambaNova Dataflow (Available for Allocation Requests)

SambaNova DataScale SN30

  • System Size: 64 Accelerators (8 nodes and 8 accelerators per node) 
  • Compute Units per Accelerator: 1280 Programmable compute units
  • Performance of a single accelerator (TFlops): >660 (BF16)
  • Software Stack Support: SambaFlow, Pytorch
  • Interconnect: Ethernet-based
SambaNova DataScale SN30 Gen 2

Graphcore Bow Pod64 (Available for Allocation Requests)

Graphcore Intelligent Processing Unit (IPU)

  • System Size: 64 Accelerators (4 nodes x 16 Accelerators per node)
  • Compute Units per Accelerator: 1472 independent processing units
  • Performance of a single accelerator (TFlops): >250 (FP16)
  • Software Stack Support: PopArt, Tensorflow, Pytorch, ONNX
  • Interconnect: IPU Link
Graphcore Bow Pod64

Habana Gaudi-1

Habana Gaudi Tensor Processing Cores

  • System Size: 16 Accelerators (2 nodes x 8 Accelerators per node)
  • Compute Units per Accelerator: 8 TPC + GEMM engine
  • Performance of a single accelerator (TFlops): >150 (FP16)
  • Software Stack Support: Synapse AI, TensorFlow and PyTorch
  • Interconnect: Ethernet-based
Habana Gaudi

Access and Support

Our AI accelerator systems are available to the research community with proven problems to solve. Researchers can now submit project proposals for the Groq, Graphcore, Cerebras CS-2 and SambaNova DataScale platforms via the ALCF’s Director’s Discretionary program. There is no submission deadline; proposals will be accepted throughout the year. Access to additional testbed resources, including Graphcore, Groq, and Habana accelerators, will be announced at a later date. 

To apply for time, submit a proposal: Allocation Request Form

More about how to use our AI Testbed: AI Testbed User Guides

Learn how to leverage our AI Testbed: AI Testbed Training

Please contact support@alcf.anl.gov with any questions.

Success Stories

Argonne Deploys New Groq System to ALCF AI Testbed

Argonne staff discuss how they use GroqRack system in the science campaigns.

Accelerating Scientific Applications With SambaNova Reconfig

Scientists at ALCF are exploring novel AI-hardware systems, such as SambaNova, in an attempt to address the challenges in scaling the performance of AI models. [Video]

Argonne collaborations bring computational tools to the forefront of COVID-19 research

The team achieved this milestone by coupling two distinct hardware platforms: Cerebras CS-1, a processor-packed silicon wafer deep learning accelerator; and ThetaGPU, an AI- and simulation-enabled extension of the Theta supercomputer with NVIDIA A100 GPUs.