Dev reverse-engineers DLSS 5 neural rendering, reimplements it bit-exact in Vulkan at 7.8ms/1080p
bdsqlsz · x · 2026-09-22
Developer maanHimself open-sourced OpenDLSS-NR, a Vulkan reimplementation of NVIDIA's DLSS 5 neural rendering network that is bit-exact against the original.
- A U-net of shifted-window transformer blocks with a global ViT: 71 blocks over six pooling levels, matching DLSS-NR build 310.8.0
- FP8 (E4M3) activations with FP16 accumulation on tensor cores, only 141 MiB of weights
- All 75 block-boundary intermediates match byte for byte, not just the final image
- Runs at 7.8ms per 1080p frame on an RTX 4070 Super, same input and output as NVIDIA's DLSS 5
- Includes an independent browser WebGPU port that runs the same bytes without tensor cores or FP8
- Users must supply their own weights; code and docs are on GitHub
More from Infra
- 63% of Americans oppose data centers in their community, spanning both parties, pollster says — AlexTensor · 2026-09-22
- Toby Ord estimates $20M spent on AI's millennium prize result, $200M for solid data — tobyordoxford · 2026-09-22
- What engineers check before adding a new LLM provider — Rama_Surasani_ · 2026-09-22
- M5 Ultra Hits 3740 tok/s Prefill on Qwen, Nearly Double Overnight — EAccelerate_42 · 2026-09-22
- Running 2.78T-param Kimi K3 on a single CPU in 8 GB RAM, no framework — tom_doerr · 2026-09-22
- LinkedIn's GPU Pre-Ranker Fuses Graph Features for 50x Bigger Models at 120ms p99 — _reachsumit · 2026-09-22