HPC, GPU & Exascale Readiness
Scientific codes written for yesterday’s CPUs leave most of today’s GPU supercomputers idle. We port, profile and scale research codes, and help groups use national supercomputing allocations and cloud HPC efficiently.
Sound familiar?
- ●Your code does not scale beyond a few nodes.
- ●You have GPU allocations but your code is CPU-only.
- ●Simulations take weeks and block your research.
- ●Your department wants a cluster but does not know what to buy.
What we offer
GPU porting
CUDA, HIP, SYCL, Kokkos and OpenMP-offload ports of research codes.
Profiling & scaling
MPI/OpenMP optimisation, I/O tuning and strong/weak scaling studies for allocation proposals.
Cloud HPC on demand
Burst-to-cloud clusters on AWS, Azure or GCP with cost controls — pay only when you compute.
Cluster design for departments
Vendor-neutral specifications, procurement support and Slurm/software-stack setup for university clusters.
Supercomputing allocation support
Scaling data and technical sections for national supercomputing (e.g. NSM/PARAM) and international allocation proposals.
Frequently asked
Do we need to share our source code?+
Only under NDA, and only the parts needed. We can also work directly on your systems.
Are you tied to a hardware vendor?+
No. Our recommendations are vendor-neutral.
Related services
Research Software Engineering & Reproducibility
Modernise research codes, make results reproducible and meet the open-science requirements of journals and funders.
Learn morePIC & Plasma Simulation with Independent V&V
Particle-in-cell, PIC-MCC and hybrid simulations — every result benchmarked against theory.
Learn moreLaser–Plasma Interaction Studies
Wakefield acceleration, ion acceleration, radiation sources and experiment design for high-power laser facilities.
Learn moreHave a simulation problem? Let’s scope it.
Tell us about your project. You will hear back from a founder within one working day, with a clear plan and a fixed quote.