New GPU Cluster Now Operational

DISAL's new high-performance computing cluster — comprising 128 NVIDIA H100 GPUs interconnected via 400 Gbps InfiniBand — is now fully operational. The system delivers over 3.2 exaFLOPS of AI compute, a 12× increase over our previous infrastructure.

The cluster is equipped with 10 PB of NVMe storage and a dedicated data pipeline for pre-processing large-scale datasets including satellite imagery, genomic sequences, and multi-hospital electronic health records.

All DISAL researchers and affiliated PhD students have immediate access via our job scheduler. External collaborators may apply for compute allocations through our research partnership programme. The upgrade was made possible through a joint investment by the University of Ghana and the NSF swarm intelligence grant.