Three tools built from our verified benchmark data and live AWS pricing. No signup, no email gate — just numbers. Useful whether you end up using Clusterra or not.
Pick a workflow (AlphaDIA, Boltz-2, OpenFE RBFE, 30× WGS, RELION cryo-EM, GROMACS…), enter a sample count, get the AWS spot bill and runtime. Numbers come from 12 verified Clusterra runs on public datasets.
Spot vs on-demand for T4, A10G, L4, L40S, V100, A100, H100, H200 across us-east-1 AZs. Filter by GPU/CPU/memory, see savings &% and AZ. Snapshot regenerated from AWS spot price history API.
Real production VRAM minimums for AlphaFold, Boltz-2, OpenFold3, RFdiffusion, RELION, GROMACS, OpenFE, and 25+ others — not the README’s toy minimums. Scaling behavior, common OOMs, recommended AWS GPUs.
Most biotech compute pages quote on-demand pricing and toy-example VRAM. Real production runs on AWS spot and real production inputs need more memory than the README suggests. We built these tools to bridge the gap — they expose the same data our customers use to scope their compute budgets.
Each tool is built from verified benchmark runs on public reference datasets (PXD, EMPIAR, GIAB, ChEMBL, PDB), with hardware, container digests, and output checksums published on the blog. The numbers are reproducible — we link to the source post on every row.
Tools we’re working on: per-tool cost calculators (Boltz-2, AlphaFold 3, OpenFold3), an sbatch generator for biotech workflows, and a public-dataset reprocess cost estimator (paste a PXD or EMPIAR ID, get the bill). If there’s a tool you wish existed, tell us.