Unique Situations need Unique Solutions
Maximize your return of investment with our services and solutions in HPC and AI-ready platforms.
Our Projects
Explore our expertise in high-performance computing and consulting services.
GPU Code Optimization
Accelerating your scientific workflows at multiple levels: 1) increasing the aggregate throughput via multiprocessing at the application level, 2) distributing your application workload on shared and distributed systems via MPI and multi-threading, and 3) porting bottlenecks to accelerators via CUDA, SYCL, OpenCL, and OpenMP.
Our expertise spans codebases such as LAMMPS, GROMACS, HOOMD-Blue, OpenMM and custom scientific codes.
The plot below shows the strong scaling performance of our LAMMPS GPU codes is close to linear and is consistently faster than the version without GPU by a factor of 3 up to 128 nodes on a world-class supercomputer.
Computing Platform Assessment
Our customers are research computing centers at universities, CIO offices, core facilities, departments starting AI initiatives and startups in materials discovery and biotech. Our projects are at multiple levels:
Develop and assess strategic plans for HPC infrastructure including server configurations, GPUs, network and storage
Optimize performance for HPC and AI workloads
Integrate cloud solutions into HPC workflows
© 2024-2026. All rights reserved.
support@codice-dev.com