cuPotential MD benchmark: 3-4x faster than stock OpenMM on an RTX 4090
gdefabritiis · x · 2026-11-17
Running benchmarks with NVIDIA cuPotential on ACEMD (now built on OpenMM), the author reports at 4 fs timestep on a 4090: dhfr single precision jumps from 2,139 to 8,151 ns/day (3.8x), with tyk2, bace variants, water100k and apo1 systems seeing consistent 3-4x speedups in both single and mixed precision. Full benchmark table published.
More from Infra
- Fact-checking the NVDA FUD: unopened Indonesia racks, $758B 2028 capex still intact — basedjensen · 2026-11-17
- Semantic Cache Pitfalls: Should LLM Caching Drop Dense Embeddings Entirely? — Artyrred · 2026-11-17
- NVIDIA cuPotential speeds up molecular dynamics 2-4x on existing GPUs without hardware switch — gdefabritiis · 2026-11-17
- A 10-item beginner reading list for LLM distributed training, from Horovod to FSDP — East-Muffin-6472 · 2026-11-17
- 18MB browser-run TTS model tiny-voca makes paid English TTS look silly — thisdudelikesAI · 2026-11-17
- Replicate quietly switches nano banana outputs from JPG downloads to presigned URLs — Kyrannio · 2026-11-17