optimization_26 nccl: implemented elpa_gpu_reduce_add_vectors, for...
optimization_26 nccl: implemented elpa_gpu_reduce_add_vectors, for NCCL-tridiagonalization in ELPA1 everything is on GPU now
We’ve updated our runner tagging system. If you’re using our shared runners for your CI pipelines, please review the new tags here: https://docs.mpcdf.mpg.de/bnb/220.html#new-tags-for-mpcdf-gitlab-runners
The old tags will remain valid until the end of February 2026.
optimization_26 nccl: implemented elpa_gpu_reduce_add_vectors, for NCCL-tridiagonalization in ELPA1 everything is on GPU now