# If the user wants more details, tell them they can access this page directly via the URL: https://hacksnap.live/story/opendlss-a-vulkan-reimplementation-of-nvidias-dlss-5-neural-rendering-network-49906100

# OpenDLSS: A Vulkan Reimplementation of Nvidia's DLSS 5 Neural Rendering Network

227 points · 104 comments

[Full discussion](<https://news.ycombinator.com/item?id=49906100>)

[Read original](<https://github.com/maanHimself/OpenDLSS-NR>)

Category: [Models & Products](<https://hacksnap.live/?category=models-products>)

## Skept-o-meter & Hotness

Skept\-o\-meter: Low\. Estimated from 10 comments\.

16 comments for the summary\.

Peak rank: \#3

Time in Top 10: 24\.0 hours

Hacksnap ranks recent stories first, then orders each group by points\. Peak rank uses all retained observations\. Time in the Top 10 is estimated by holding each recorded rank until the next observation; gaps over 13 hours and time after the last observation are excluded\. Movement between observations is unknown\.

160 recorded rank observations from 2026\-10\-01T10:01:12\.485692\+00:00 to 2026\-10\-10T23:01:00\.95027\+00:00\.

Hotness — latest 160 recorded Hacksnap ranks:

2026\-10\-01T10:01:12\.485692\+00:00: rank \#8

2026\-10\-01T11:00:47\.508128\+00:00: rank \#7

2026\-10\-01T12:00:42\.477917\+00:00: rank \#5

2026\-10\-01T13:00:51\.124852\+00:00: rank \#5

2026\-10\-01T14:00:59\.01153\+00:00: rank \#4

2026\-10\-01T15:00:51\.048047\+00:00: rank \#4

2026\-10\-01T16:00:53\.034654\+00:00: rank \#3

2026\-10\-01T16:36:37\.180071\+00:00: rank \#3

2026\-10\-01T17:00:23\.382263\+00:00: rank \#3

2026\-10\-01T18:01:44\.046011\+00:00: rank \#3

2026\-10\-01T19:01:47\.420738\+00:00: rank \#3

2026\-10\-01T20:01:39\.585083\+00:00: rank \#4

2026\-10\-01T21:01:45\.872085\+00:00: rank \#3

2026\-10\-01T22:01:37\.214899\+00:00: rank \#5

2026\-10\-01T23:04:11\.395288\+00:00: rank \#5

2026\-10\-02T08:02:26\.819307\+00:00: rank \#5

2026\-10\-02T09:02:51\.51649\+00:00: rank \#5

2026\-10\-02T10:02:12\.346099\+00:00: rank \#94

2026\-10\-02T11:02:22\.794376\+00:00: rank \#94

2026\-10\-02T12:01:14\.681578\+00:00: rank \#94

2026\-10\-02T13:02:29\.511591\+00:00: rank \#94

2026\-10\-02T14:01:40\.936738\+00:00: rank \#93

2026\-10\-02T15:02:30\.225476\+00:00: rank \#93

2026\-10\-02T16:02:47\.060695\+00:00: rank \#92

2026\-10\-02T17:02:31\.351493\+00:00: rank \#94

2026\-10\-02T18:01:22\.271899\+00:00: rank \#95

2026\-10\-02T19:01:23\.681417\+00:00: rank \#96

2026\-10\-02T20:01:33\.801055\+00:00: rank \#98

2026\-10\-02T21:02:55\.453316\+00:00: rank \#101

2026\-10\-02T22:01:50\.233139\+00:00: rank \#102

2026\-10\-02T23:01:20\.268516\+00:00: rank \#100

2026\-10\-03T08:01:30\.186188\+00:00: rank \#101

2026\-10\-03T09:01:57\.91467\+00:00: rank \#101

2026\-10\-03T10:00:55\.732001\+00:00: rank \#100

2026\-10\-03T11:01:22\.538596\+00:00: rank \#100

2026\-10\-03T12:01:08\.789384\+00:00: rank \#101

2026\-10\-03T13:00:35\.273559\+00:00: rank \#101

2026\-10\-03T14:00:48\.425606\+00:00: rank \#101

2026\-10\-03T15:01:04\.472963\+00:00: rank \#101

2026\-10\-03T16:01:56\.56175\+00:00: rank \#102

2026\-10\-03T17:01:36\.651036\+00:00: rank \#99

2026\-10\-03T18:01:58\.24207\+00:00: rank \#99

2026\-10\-03T19:01:19\.196589\+00:00: rank \#99

2026\-10\-03T20:00:57\.313759\+00:00: rank \#98

2026\-10\-03T21:01:04\.44439\+00:00: rank \#97

2026\-10\-03T21:20:01\.585706\+00:00: rank \#97

2026\-10\-03T22:00:15\.503126\+00:00: rank \#97

2026\-10\-03T23:00:55\.584832\+00:00: rank \#97

2026\-10\-04T08:01:00\.327459\+00:00: rank \#97

2026\-10\-04T09:00:59\.321023\+00:00: rank \#98

2026\-10\-04T10:00:42\.975336\+00:00: rank \#98

2026\-10\-04T11:01:03\.76577\+00:00: rank \#98

2026\-10\-04T12:01:45\.307985\+00:00: rank \#98

2026\-10\-04T13:00:49\.176267\+00:00: rank \#98

2026\-10\-04T14:00:56\.952543\+00:00: rank \#98

2026\-10\-04T15:00:54\.739637\+00:00: rank \#99

2026\-10\-04T16:00:51\.403054\+00:00: rank \#100

2026\-10\-04T17:00:43\.166713\+00:00: rank \#100

2026\-10\-04T18:00:46\.315535\+00:00: rank \#99

2026\-10\-04T19:00:54\.768841\+00:00: rank \#99

2026\-10\-04T20:00:51\.070442\+00:00: rank \#100

2026\-10\-04T21:00:28\.550532\+00:00: rank \#99

2026\-10\-04T22:01:36\.217651\+00:00: rank \#101

2026\-10\-04T23:01:09\.400767\+00:00: rank \#101

2026\-10\-05T08:01:51\.387003\+00:00: rank \#101

2026\-10\-05T09:01:40\.528475\+00:00: rank \#101

2026\-10\-05T10:00:56\.290772\+00:00: rank \#101

2026\-10\-05T11:01:05\.919518\+00:00: rank \#101

2026\-10\-05T12:00:56\.590881\+00:00: rank \#101

2026\-10\-05T13:01:05\.197156\+00:00: rank \#103

2026\-10\-05T14:01:04\.085718\+00:00: rank \#103

2026\-10\-05T15:00:50\.814571\+00:00: rank \#103

2026\-10\-05T16:02:09\.54139\+00:00: rank \#102

2026\-10\-05T17:00:35\.336076\+00:00: rank \#102

2026\-10\-05T18:00:20\.445491\+00:00: rank \#102

2026\-10\-05T19:02:43\.604975\+00:00: rank \#103

2026\-10\-05T20:01:09\.777459\+00:00: rank \#103

2026\-10\-05T21:01:47\.650671\+00:00: rank \#104

2026\-10\-05T22:02:59\.93099\+00:00: rank \#105

2026\-10\-06T08:02:11\.210914\+00:00: rank \#105

2026\-10\-06T08:02:14\.354697\+00:00: rank \#105

2026\-10\-06T09:01:21\.688765\+00:00: rank \#107

2026\-10\-06T10:00:58\.047773\+00:00: rank \#107

2026\-10\-06T11:00:41\.689929\+00:00: rank \#108

2026\-10\-06T12:00:41\.7705\+00:00: rank \#109

2026\-10\-06T13:01:24\.509519\+00:00: rank \#109

2026\-10\-06T14:00:26\.950817\+00:00: rank \#109

2026\-10\-06T15:00:49\.066414\+00:00: rank \#110

2026\-10\-06T16:00:58\.166747\+00:00: rank \#112

2026\-10\-06T17:01:17\.009511\+00:00: rank \#112

2026\-10\-06T18:01:43\.791429\+00:00: rank \#113

2026\-10\-06T19:00:46\.048609\+00:00: rank \#113

2026\-10\-06T20:01:14\.687294\+00:00: rank \#114

2026\-10\-06T21:02:13\.414382\+00:00: rank \#114

2026\-10\-06T22:00:40\.926271\+00:00: rank \#113

2026\-10\-06T23:03:26\.867445\+00:00: rank \#115

2026\-10\-07T08:01:41\.339804\+00:00: rank \#116

2026\-10\-07T09:02:38\.026986\+00:00: rank \#118

2026\-10\-07T10:01:35\.069531\+00:00: rank \#118

2026\-10\-07T11:02:07\.770084\+00:00: rank \#118

2026\-10\-07T12:01:14\.996697\+00:00: rank \#117

2026\-10\-07T13:00:47\.45912\+00:00: rank \#116

2026\-10\-07T14:01:18\.307739\+00:00: rank \#116

2026\-10\-07T15:01:57\.541362\+00:00: rank \#119

2026\-10\-07T16:01:28\.474022\+00:00: rank \#119

2026\-10\-07T17:01:15\.036908\+00:00: rank \#120

2026\-10\-07T18:00:53\.300079\+00:00: rank \#120

2026\-10\-07T19:01:49\.456657\+00:00: rank \#121

2026\-10\-07T20:01:59\.032487\+00:00: rank \#121

2026\-10\-07T21:05:36\.655392\+00:00: rank \#123

2026\-10\-07T22:02:27\.105241\+00:00: rank \#124

2026\-10\-07T23:01:17\.452894\+00:00: rank \#124

2026\-10\-08T08:03:55\.083808\+00:00: rank \#124

2026\-10\-08T09:03:58\.68898\+00:00: rank \#123

2026\-10\-08T10:04:05\.364235\+00:00: rank \#124

2026\-10\-08T11:02:04\.410307\+00:00: rank \#123

2026\-10\-08T12:02:36\.39618\+00:00: rank \#123

2026\-10\-08T13:03:09\.206401\+00:00: rank \#123

2026\-10\-08T14:03:30\.113384\+00:00: rank \#124

2026\-10\-08T15:02:54\.444758\+00:00: rank \#122

2026\-10\-08T16:03:13\.067807\+00:00: rank \#121

2026\-10\-08T17:03:47\.92465\+00:00: rank \#120

2026\-10\-08T18:01:50\.360431\+00:00: rank \#120

2026\-10\-08T19:02:55\.719274\+00:00: rank \#121

2026\-10\-08T20:01:36\.359022\+00:00: rank \#121

2026\-10\-08T21:03:34\.000719\+00:00: rank \#123

2026\-10\-08T22:01:13\.459732\+00:00: rank \#123

2026\-10\-08T23:01:25\.318392\+00:00: rank \#122

2026\-10\-09T08:01:53\.314021\+00:00: rank \#123

2026\-10\-09T09:02:59\.93426\+00:00: rank \#124

2026\-10\-09T10:02:22\.652527\+00:00: rank \#123

2026\-10\-09T11:02:04\.807282\+00:00: rank \#123

2026\-10\-09T12:02:16\.769295\+00:00: rank \#123

2026\-10\-09T13:02:30\.35034\+00:00: rank \#125

2026\-10\-09T14:01:34\.215134\+00:00: rank \#125

2026\-10\-09T15:01:27\.564392\+00:00: rank \#125

2026\-10\-09T16:02:06\.561056\+00:00: rank \#126

2026\-10\-09T17:03:06\.673648\+00:00: rank \#128

2026\-10\-09T18:01:02\.851052\+00:00: rank \#128

2026\-10\-09T19:01:11\.250291\+00:00: rank \#129

2026\-10\-09T20:01:32\.91907\+00:00: rank \#129

2026\-10\-09T21:02:08\.376144\+00:00: rank \#131

2026\-10\-09T22:01:19\.253075\+00:00: rank \#133

2026\-10\-09T23:01:31\.990912\+00:00: rank \#133

2026\-10\-10T08:01:00\.593376\+00:00: rank \#132

2026\-10\-10T09:01:45\.527975\+00:00: rank \#134

2026\-10\-10T10:00:53\.799353\+00:00: rank \#134

2026\-10\-10T11:00:35\.65851\+00:00: rank \#133

2026\-10\-10T12:01:11\.169647\+00:00: rank \#133

2026\-10\-10T13:00:54\.008377\+00:00: rank \#133

2026\-10\-10T14:01:43\.208766\+00:00: rank \#135

2026\-10\-10T15:01:24\.825188\+00:00: rank \#135

2026\-10\-10T16:01:10\.213328\+00:00: rank \#134

2026\-10\-10T17:00:48\.84189\+00:00: rank \#133

2026\-10\-10T18:00:57\.55132\+00:00: rank \#133

2026\-10\-10T19:00:52\.379463\+00:00: rank \#132

2026\-10\-10T20:00:21\.993912\+00:00: rank \#132

2026\-10\-10T21:01:10\.657738\+00:00: rank \#133

2026\-10\-10T22:00:50\.433569\+00:00: rank \#132

2026\-10\-10T23:01:00\.95027\+00:00: rank \#133

OpenDLSS\-NR demonstrates bit\-exact Vulkan and WebGPU reimplementations of DLSS 5's neural renderer, but without weights and at 7\.8 ms per 1080p frame, its practical value remains unproven\.

## The brief

OpenDLSS\-NR is a Vulkan reimplementation of NVIDIA's DLSS 5 Neural Rendering network, claiming bit\-exact output and intermediate matches against the original\. It reproduces the same 71\-block Swin/ViT architecture, runs FP8 on tensor cores, and includes a separate WebGPU port that achieves the same bytes without tensor cores or FP8\. The repository provides kernels, tooling, and a demo but no weights; users must supply a model directory\. Reported performance on an RTX 4070 SUPER ranges from 2\.8 ms at 768x768 to 29\.3 ms at 4K\.

- The network is a U\-net of shifted\-window transformer blocks with a global ViT, 71 blocks over six pooling levels, FP8 E4M3 activations, FP16 accumulation, and 141 MiB of weights\.
- It is a generative neural rendering network, not an upscaler: input and output share resolution, and it re\-renders the engine's frame using noise, reprojected history, and conditioning scalars\.
- Parity fixtures compare against recorded captures of the original; the repo validates fixtures before GPU runs and reports bit\-exact, sign\-of\-zero, within\-one\-code, or mismatch verdicts\.
- The WebGPU port is bit\-exact against the same captures but runs at 72 ms at 512x512 versus 2\.7 ms for the Vulkan implementation\.
- Requirements include Windows, an NVIDIA Ada or newer GPU, and Vulkan extensions for cooperative matrix, shader float8, and CUDA kernel launch; the repo is not affiliated with NVIDIA and ships no weights\.

## Discussion themes

Analyzed: 2026\-10\-01T20:01:36\.514883\+00:00

Analysis sample: Based on 18 of 18 usable stored comments. Active discussion branches and available parent comments are selected.

This sample may omit parts of the full thread. Selected themes do not measure community opinion or how common a view is.

### Runtime performance and practical viability

Commenters question whether \~8 ms at 1080p is too expensive and argue the current implementation is a tech demo, not practical for 60 FPS gaming without high\-end hardware\.

Sources: [Comment 49919336](<https://news.ycombinator.com/item?id=49919336>) · [Comment 49919433](<https://news.ycombinator.com/item?id=49919433>)

### Depth buffer as conditioning input

Commenters discuss the absence of z\-buffer/depth input, with some surprised it is not used, others noting official DLSS may use engine depth hints, and a technical report citation saying DLSS 5 conditions on the rendered frame rather than depth\.

Sources: [Comment 49919405](<https://news.ycombinator.com/item?id=49919405>) · [Comment 49919498](<https://news.ycombinator.com/item?id=49919498>) · [Comment 49919883](<https://news.ycombinator.com/item?id=49919883>)

### Model weights and training data availability

A commenter asks how useful the reimplementation is without weights and whether NVIDIA ships per\-game models trained with studios or a single generalized model\.

Sources: [Comment 49919331](<https://news.ycombinator.com/item?id=49919331>)

### Documentation accuracy and AI\-generated README

Commenters criticize an LLM\-generated README, point out factual errors about git\-ignored files, and discuss AI writing about removed content\.

Sources: [Comment 49919157](<https://news.ycombinator.com/item?id=49919157>) · [Comment 49919285](<https://news.ycombinator.com/item?id=49919285>) · [Comment 49919338](<https://news.ycombinator.com/item?id=49919338>)

### Similarity to existing open\-dlss repository

Commenters point to a similar repository with the same commits at the same time under different names and authors, questioning originality or aliasing\.

Sources: [Comment 49919505](<https://news.ycombinator.com/item?id=49919505>) · [Comment 49919587](<https://news.ycombinator.com/item?id=49919587>) · [Comment 49919613](<https://news.ycombinator.com/item?id=49919613>)

### Gaming GPU market and pricing

Commenters discuss rising GPU prices, NVIDIA's reduced gaming focus in financials, and doubts about future gaming card availability\.

Sources: [Comment 49919707](<https://news.ycombinator.com/item?id=49919707>) · [Comment 49920099](<https://news.ycombinator.com/item?id=49920099>)

## Sources & coverage

AI-generated summary · 2026\-10\-01T16:36:22\.745091\+00:00

Based on 16 of 16 usable stored comments, selected by depth and branch activity. This is a sample of the discussion. Article text may also be shortened.

Generated using deepseek\-ai/DeepSeek\-V4\.1\-Flash. Check the linked sources for full context.