Opens in a new tab

Debugging & Profiling for HPC and AI

Proven on the World's Supercomputers.
Now Optimized for AI Infrastructure.

Find bugs and performance bottlenecks in parallel code, from one node to thousands.

Trusted by leading supercomputing centres and research institutes

One Toolset for HPC and AI

The same debugger, profiler and performance reports serve supercomputing centres and GPU clusters. Choose your path.

HPC & Scientific Computing

Simulation and Research Codes at Scale

MPI and OpenMP applications in C, C++, Fortran and Python, from a single node to thousands, on x86-64 and Arm CPUs with NVIDIA and AMD GPUs.

AI Infrastructure

GPU Clusters for Training and Inference

Distributed GPU workloads with CUDA and ROCm code and NCCL communication. New in Linaro Forge 26.0: profile NCCL collectives with Linaro MAP and Performance Reports.

Why Linaro Forge

Battle-tested for more than twenty years by the world’s most demanding computing centres.

Extreme Scalability

Debug and profile where other tools can’t go: hundreds of thousands of processes in one session.

Deep Cross-Platform Support

Years of partnership with hardware vendors, from x86-64 and Arm CPUs to NVIDIA and AMD GPUs.

Built for Domain Experts

Powerful tools that scientists and model builders can use, without being computer scientists.

See why teams choose Linaro Forge

Compute Loss Calculator

Estimate what crashed and hung jobs cost each year, whether your cluster is counted in GPUs or CPU nodes.

Your cloud price or internal charge-back rate.

Estimated annual wasted
compute spend

$1,032,192

Annual savings if Forge prevents 80% of these failures

$825,754

Schedule ROI Review

Try Linaro Forge on Your Cluster

Try Linaro Forge free for two weeks on up to 32 MPI ranks.

Start a Free Trial