Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

🧵(3/4) 📜 Paper 3 - DGF: A Dense, Hardware-Friendly Geometry Format for Lossily Compressing Meshlets with Arbitrary Topologies. DGF will do for geometry what block compression did for textures, enabling #raytracing of massive micropolygon geometry 🗜️

AMD GPUOpen

18,672 subscribers

14,201 views • 2 years ago •via X (Twitter)

Gaming Science & Technology #raytracing

Anya Rossi• Live Now

Private livecam show

4 Comments

AMD GPUOpen2 years ago

🧵(1/4) We'll be at @HPG_Conf, Denver, July 27! We're very excited to share the four papers we'll be presenting in this thread 🔥 📜 Paper 1 - HIPRT: Our #raytracing framework in HIP Take a look (28MB): Learn more about #HIP RT:

AMD GPUOpen2 years ago

🧵(2/4) @HPG_Conf 📜 Paper 2 - H-PLOC: Super-fast, high-quality BVH construction. Allows for constructing a BVH over millions of triangles in just a few ms. Very useful for efficiently #raytracing a large amount of fully dynamic geometry per frame 🎇

AMD GPUOpen2 years ago

🧵(4/4) @HPG_Conf 📜 Paper 4 - Real-Time Procedural Generation with GPU Work Graphs This paper shows some of the procedural effects from our #GDC2024 #WorkGraphs demo! 🏰 (32MB) Demo:

Brian Karis2 years ago

Very happy to see this published! DGF is very similar to Nanite's encoding minus attributes. Switching Nanite to this format would be some work but is definitely possible. Doing so wouldn't be much of a win for rasterization but the really exciting part here is ray tracing it.

Related Videos

Really proud of what I was able to achieve developing 3D cell nuclei with custom geometry shaders using Three.js for a cancer care platform, all built with Google Gemini 3 Pro AI 💜 Had to approach the project twice, but had a lot of fun doing it :) Live:

Really proud of what I was able to achieve developing 3D cell nuclei with custom geometry shaders using Three.js for a cancer care platform, all built with Google Gemini 3 Pro AI 💜 Had to approach the project twice, but had a lot of fun doing it :) Live:

The Bugged Dev

19,510 views • 7 months ago

Might not be a bad idea for a barbershop website, huh? :D Vibe coded a paper-fold image effect and 3D point-cloud text geometry using custom GLSL shaders in Three.js with Gemini 3 Pro AI all in Google AI Studio. It was soo soo fun!! Live: Code:

Might not be a bad idea for a barbershop website, huh? :D Vibe coded a paper-fold image effect and 3D point-cloud text geometry using custom GLSL shaders in Three.js with Gemini 3 Pro AI all in Google AI Studio. It was soo soo fun!! Live: Code:

The Bugged Dev

20,854 views • 6 months ago

Did you know that in Blender you can easily make insets with the 'Intersect Knife' operator? ✨ 1. Select the geometry you want to cut out 2. Press Ctrl+F,K 3. Remove the original geometry and select the newly cut faces 4. Press Ctrl+F,F to extrude along normals Important: when you're extruding faces along normals, be sure to also press S after you press Ctrl+F,F to enable the "Even Thickness" option for a better result! So the full shortcut is Ctrl+F,F,S (yes, really) #b3d

Did you know that in Blender you can easily make insets with the 'Intersect Knife' operator? ✨ 1. Select the geometry you want to cut out 2. Press Ctrl+F,K 3. Remove the original geometry and select the newly cut faces 4. Press Ctrl+F,F to extrude along normals Important: when you're extruding faces along normals, be sure to also press S after you press Ctrl+F,F to enable the "Even Thickness" option for a better result! So the full shortcut is Ctrl+F,F,S (yes, really) #b3d

passivestar

290,164 views • 3 years ago

To support massive, high-fidelity worlds, our engine implements an efficient LOD strategy with a cloud transcoding system that streams meshes and textures, even on lower end mobile devices. We are expanding this framework to AI upsampling of 3D worlds from prompts. In this preview video below, with a single user prompt, our 3D upsampling technology updates the geometry and texture to transform the classic Roblox game Crossroads into a fantasy world with rich organic detail. 3/4

To support massive, high-fidelity worlds, our engine implements an efficient LOD strategy with a cloud transcoding system that streams meshes and textures, even on lower end mobile devices. We are expanding this framework to AI upsampling of 3D worlds from prompts. In this preview video below, with a single user prompt, our 3D upsampling technology updates the geometry and texture to transform the classic Roblox game Crossroads into a fantasy world with rich organic detail. 3/4

Roblox

1,422,335 views • 6 months ago

Claude Fable 5 has been out for a day. I already one-shotted a voxel spaceship racer with it. I fed it a reference image of a track layout and told it to build a full game around that shape. Missiles, boost zones, 3 AI opponents, 3 laps, a HUD with speed and health. Fable 5 read the image, traced the track geometry, and built the entire thing. One prompt. This is what Mythos-class models do to the build cycle.

Claude Fable 5 has been out for a day. I already one-shotted a voxel spaceship racer with it. I fed it a reference image of a track layout and told it to build a full game around that shape. Missiles, boost zones, 3 AI opponents, 3 laps, a HUD with speed and health. Fable 5 read the image, traced the track geometry, and built the entire thing. One prompt. This is what Mythos-class models do to the build cycle.

God of Prompt

27,386 views • 2 months ago

How to convert a paper into a poster in seconds? 1. Go to 2. Upload the PDF of your research paper. 3. Enter the prompt to specify your poster requirements 4. That’s it. SciSpace will create a perfect poster for you 5. Once ready, download your poster in PPT format

How to convert a paper into a poster in seconds? 1. Go to 2. Upload the PDF of your research paper. 3. Enter the prompt to specify your poster requirements 4. That’s it. SciSpace will create a perfect poster for you 5. Once ready, download your poster in PPT format

Faheem Ullah

22,696 views • 8 months ago

Holy moly, I finally vibe coded a website I’ve always dreamt of building 🥹 Nah, it’s not a Blender model. Not even for 3D podium. It’s pure geometry built using Three.js with color maps (albedo), spring-based physics, a hybrid DOM, real-time planar reflections, and more. Completely vibe coded all using Gemini 3 Pro in Google AI Studio!! Live:

Holy moly, I finally vibe coded a website I’ve always dreamt of building 🥹 Nah, it’s not a Blender model. Not even for 3D podium. It’s pure geometry built using Three.js with color maps (albedo), spring-based physics, a hybrid DOM, real-time planar reflections, and more. Completely vibe coded all using Gemini 3 Pro in Google AI Studio!! Live:

The Bugged Dev

184,472 views • 6 months ago

Geometry 🤝 Topology What you're seeing here is called Hopf Fibration. In pure-math terms, S³ (the unit sphere in 4D) can be decomposed into circles called fibers. The Hopf map sends each of those circles to a single point on S². So you don’t fill S³ with random loops, you fill it with a perfectly organized family of circles, one circle sitting above each point of S². For the animation, we sample thousands of fibers on S³, apply a smooth SO(4) rotation (a real rotation in 4D, not a 3D cheat), then stereographically project into R³ so we can actually draw it. Pure mathematics is beautiful alz #HopfFibration #Topology #Geometry #StereographicProjection #PureMathematics

Geometry 🤝 Topology What you're seeing here is called Hopf Fibration. In pure-math terms, S³ (the unit sphere in 4D) can be decomposed into circles called fibers. The Hopf map sends each of those circles to a single point on S². So you don’t fill S³ with random loops, you fill it with a perfectly organized family of circles, one circle sitting above each point of S². For the animation, we sample thousands of fibers on S³, apply a smooth SO(4) rotation (a real rotation in 4D, not a 3D cheat), then stereographically project into R³ so we can actually draw it. Pure mathematics is beautiful alz #HopfFibration #Topology #Geometry #StereographicProjection #PureMathematics

Mathelirium

15,727 views • 6 months ago

Microsoft open-sourced a 4B model that turns any image into a production-ready 3D asset in 3 seconds. It’s called TRELLIS.2, a fully textured, physically accurate 3D models with PBR textures out of the box. → Full PBR (base color, roughness, metallic, opacity) → Handles hair, cloth, glass, non-manifold geometry → Exports .glb ready for Unity/Unreal/Blender → Runs locally, ships in 3 seconds It's not a demo or a research preview. The full training codebase is public. You can fine-tune it on your own asset library and get a model that generates in your studio's exact style. 100% Open Source

Microsoft open-sourced a 4B model that turns any image into a production-ready 3D asset in 3 seconds. It’s called TRELLIS.2, a fully textured, physically accurate 3D models with PBR textures out of the box. → Full PBR (base color, roughness, metallic, opacity) → Handles hair, cloth, glass, non-manifold geometry → Exports .glb ready for Unity/Unreal/Blender → Runs locally, ships in 3 seconds It's not a demo or a research preview. The full training codebase is public. You can fine-tune it on your own asset library and get a model that generates in your studio's exact style. 100% Open Source

Superman

279,292 views • 12 days ago

Dear Nigerians, Bola Ahmed Tinubu has borrowed more than ₦71 trillion in 3 years. But what did he do with the money? Jude Bela deserves our massive following for always breaking the numbers down. This will break your heart. 💔

Dear Nigerians, Bola Ahmed Tinubu has borrowed more than ₦71 trillion in 3 years. But what did he do with the money? Jude Bela deserves our massive following for always breaking the numbers down. This will break your heart. 💔

NEFERTITI

92,274 views • 3 months ago

The Freemasons did not invent the title of Great Architect for God arbitrarily. Randall traces it to a specific and ancient understanding - that the universe was built according to geometry, that all of creation carries that geometry inherently within it, and that the role of the master builder was not to express personal creativity but to replicate the processes by which God constructed reality itself. The grand geometric mean - another name Randall notes the Freemasons apply to the divine - points to the same conclusion. Mathematics and divinity were never separate categories in this tradition. They were the same thing viewed from different angles. What Randall finds most compelling is the point at which observation overtakes skepticism. The geometry is there. It appears in the proportions of living organisms, in the structure of galaxies, in the dimensions of ancient monuments, and in the relationships between celestial bodies. Randall’s position is straightforward - there is a threshold beyond which dismissing this as coincidence is no longer intellectually honest. Once you begin looking for the geometry of creation, it becomes impossible to stop finding it. The ancient master builders understood that - and everything they built was a deliberate act of alignment with a geometric order they believed was not invented by human minds but discovered within a universe that was already built that way from the beginning.

The Freemasons did not invent the title of Great Architect for God arbitrarily. Randall traces it to a specific and ancient understanding - that the universe was built according to geometry, that all of creation carries that geometry inherently within it, and that the role of the master builder was not to express personal creativity but to replicate the processes by which God constructed reality itself. The grand geometric mean - another name Randall notes the Freemasons apply to the divine - points to the same conclusion. Mathematics and divinity were never separate categories in this tradition. They were the same thing viewed from different angles. What Randall finds most compelling is the point at which observation overtakes skepticism. The geometry is there. It appears in the proportions of living organisms, in the structure of galaxies, in the dimensions of ancient monuments, and in the relationships between celestial bodies. Randall’s position is straightforward - there is a threshold beyond which dismissing this as coincidence is no longer intellectually honest. Once you begin looking for the geometry of creation, it becomes impossible to stop finding it. The ancient master builders understood that - and everything they built was a deliberate act of alignment with a geometric order they believed was not invented by human minds but discovered within a universe that was already built that way from the beginning.

Randall Carlson

11,848 views • 4 months ago

Titanium is one of the hardest materials to machine. It cuts much slower than aluminum, runs hot, and deep bowl-like geometry will snap a tool if chips aren't continuously flushed out with coolant. That strength is what allows thin, complex parts to hold tight tolerances, making it the material of choice for high-performance satellites that have to survive the harshest environment there is. Astranis machines titanium in-house at our factory in Northern California, building world-class hardware on American soil.

Titanium is one of the hardest materials to machine. It cuts much slower than aluminum, runs hot, and deep bowl-like geometry will snap a tool if chips aren't continuously flushed out with coolant. That strength is what allows thin, complex parts to hold tight tolerances, making it the material of choice for high-performance satellites that have to survive the harshest environment there is. Astranis machines titanium in-house at our factory in Northern California, building world-class hardware on American soil.

Astranis Space Technologies

21,172 views • 4 months ago

Meet Tencent HY 3D Studio 1.2 👋 With this major upgrade to our 3D creation pipeline, you can generate assets with sculpt-level detail and fine-grained interactive control. Starting today, the studio is officially open for Public Beta — no application required. Tencent HY 3D Studio 1.2 introduces a suite of features built for creative precision, including: 🧩 PartGen 1.5: High-Precision Component Partitioning — 1536³ Resolution: Boosted from 1024³ for crystal-clear model splitting and ultra-fine detail retention. — Fine-Grained Interaction: Introducing intuitive, brush-based control for precise, manual component editing. — Shape Adherence: Drastically improved geometry integrity for even the most intricate objects. 🎨 Tencent HY 3D 3.1: Sculpt-Level Detail — Enhanced Geometry: Fine-grained details that adapt perfectly to stylized inputs.. — Texture Fidelity: A massive leap in color accuracy and texture fidelity, reflecting your original input with stunning realism. — 8-View Control: Expanded from 4 to 8 input views for ultimate reconstruction accuracy. ✨We’re so excited to see what you create! Try it now:

Meet Tencent HY 3D Studio 1.2 👋 With this major upgrade to our 3D creation pipeline, you can generate assets with sculpt-level detail and fine-grained interactive control. Starting today, the studio is officially open for Public Beta — no application required. Tencent HY 3D Studio 1.2 introduces a suite of features built for creative precision, including: 🧩 PartGen 1.5: High-Precision Component Partitioning — 1536³ Resolution: Boosted from 1024³ for crystal-clear model splitting and ultra-fine detail retention. — Fine-Grained Interaction: Introducing intuitive, brush-based control for precise, manual component editing. — Shape Adherence: Drastically improved geometry integrity for even the most intricate objects. 🎨 Tencent HY 3D 3.1: Sculpt-Level Detail — Enhanced Geometry: Fine-grained details that adapt perfectly to stylized inputs.. — Texture Fidelity: A massive leap in color accuracy and texture fidelity, reflecting your original input with stunning realism. — 8-View Control: Expanded from 4 to 8 input views for ultimate reconstruction accuracy. ✨We’re so excited to see what you create! Try it now:

Tencent Hy

83,314 views • 6 months ago

Elon: "There is some chance that HW3 does not achieve the safety level that allows for unsupervised FSD. There is some chance of that. And if that turns out to be the case, we will upgrade those who bought hardware 3 FSD for free. We're going to have to upgrade people's hardware 3 computer for those that have bought Full Self-Driving. And that is the honest answer. And that's going to be painful and difficult, but we'll get it done. What we want to do is we want to get unsupervised done on Hardware 4 first. Once it's done, then we'll go back and look at what we need to do with Hardware 3 cars. Yeah, we've not completely given up on HW3, but we do want to solve autonomy first, and then we'll come back with a way to take care of customers. These customers are very important. They were the early adopters. We will definitely take care of you guys. Once the V14 release series is fully done, we are planning on working on a V14 lite version for Hardware 3, probably expected in Q2 next year"

Elon: "There is some chance that HW3 does not achieve the safety level that allows for unsupervised FSD. There is some chance of that. And if that turns out to be the case, we will upgrade those who bought hardware 3 FSD for free. We're going to have to upgrade people's hardware 3 computer for those that have bought Full Self-Driving. And that is the honest answer. And that's going to be painful and difficult, but we'll get it done. What we want to do is we want to get unsupervised done on Hardware 4 first. Once it's done, then we'll go back and look at what we need to do with Hardware 3 cars. Yeah, we've not completely given up on HW3, but we do want to solve autonomy first, and then we'll come back with a way to take care of customers. These customers are very important. They were the early adopters. We will definitely take care of you guys. Once the V14 release series is fully done, we are planning on working on a V14 lite version for Hardware 3, probably expected in Q2 next year"

Whole Mars Catalog

700,623 views • 6 months ago

What would Super Mario 64 look like were it enhanced for the Sega Dreamcast? Well, modder Esppiral is setting out to answer exactly that question with his upcoming mod for our Mario 64 DC port, featuring enhanced geometry and textures, as well as Mario Galaxy's fresnel lighting! While this footage is still extremely early, and represents a work in progress, it is being brought to you by the same mad Dreamcast scientist who remastered Dead or Alive 2 and brought Xbox assets and effects to our port of Grand Theft Auto 3! Please be sure to support the original video and subscribe to Esppiral's channel for future updates and more mad Sega Dreamcast science:

What would Super Mario 64 look like were it enhanced for the Sega Dreamcast? Well, modder Esppiral is setting out to answer exactly that question with his upcoming mod for our Mario 64 DC port, featuring enhanced geometry and textures, as well as Mario Galaxy's fresnel lighting! While this footage is still extremely early, and represents a work in progress, it is being brought to you by the same mad Dreamcast scientist who remastered Dead or Alive 2 and brought Xbox assets and effects to our port of Grand Theft Auto 3! Please be sure to support the original video and subscribe to Esppiral's channel for future updates and more mad Sega Dreamcast science:

Falco Girgis

119,496 views • 1 year ago

Starting the new year without human labeling 🎉!! Multimodal lidar-camera data is a gold mine of dense 3D geometry hiding in plain sight. For supervised pretraining and validation at scale at Torc-Robotics, we rely on fully automated pseudo-labeling pipelines. Exploiting geometric priors from temporally accumulated LiDAR maps and an iterative update rule enforces joint geometric–semantic consistency while detecting moving objects via inconsistencies. We achieve 3D semantic labels and 3D bounding boxes with human-like quality at 200m+ range required for highway driving. Paper: Exciting work with Torc-Robotics with Filippo Ghilotti, Samuel Brucker, Nahku Saidy, Matteo Matteucci, Mario Bijelic.

Starting the new year without human labeling 🎉!! Multimodal lidar-camera data is a gold mine of dense 3D geometry hiding in plain sight. For supervised pretraining and validation at scale at Torc-Robotics, we rely on fully automated pseudo-labeling pipelines. Exploiting geometric priors from temporally accumulated LiDAR maps and an iterative update rule enforces joint geometric–semantic consistency while detecting moving objects via inconsistencies. We achieve 3D semantic labels and 3D bounding boxes with human-like quality at 200m+ range required for highway driving. Paper: Exciting work with Torc-Robotics with Filippo Ghilotti, Samuel Brucker, Nahku Saidy, Matteo Matteucci, Mario Bijelic.

Felix Heide

18,210 views • 7 months ago

Wow. Recreating the Shawshank Redemption prison in 3D from a single video, in real time (!) Just read the MASt3R-SLAM paper and it's pretty neat. These folks basically built a real-time dense SLAM system on top of MASt3R, which is a transformer-based neural network that can do 3d reconstruction and localization from uncalibrated image pairs. The cool part is they don't need a fixed camera model -- it just works with arbitrary cameras -- think different focal lengths, sensor sizes, even handling zooming in video (FMV drone video anyone?!). If you've done photogrammetry or played with NeRFs you know that is a HUGE deal. They've solved some tricky problems like efficient point matching and tracking, plus they've figured out how to fuse point clouds and handle loop closures in real-time. Their system runs at about 15 FPS on a 4090 and produces both camera poses and dense geometry. When they know the camera calibration, they get SOTA results across several benchmarks, but even without calibration, they still perform well. What's interesting is the approach -- most recent SLAM work has built on DROID-SLAM's architecture, but these folks went a different direction by leveraging a strong 3D reconstruction prior. Seems to give them more coherent geometry, which makes sense since that's what MASt3R was designed for. For anyone who cares about monocular SLAM and 3D reconstruction, this feels like a significant step toward plug-and-play dense SLAM without calibration headaches -- perfect for drones, robots, AR/VR -- the works!

Wow. Recreating the Shawshank Redemption prison in 3D from a single video, in real time (!) Just read the MASt3R-SLAM paper and it's pretty neat. These folks basically built a real-time dense SLAM system on top of MASt3R, which is a transformer-based neural network that can do 3d reconstruction and localization from uncalibrated image pairs. The cool part is they don't need a fixed camera model -- it just works with arbitrary cameras -- think different focal lengths, sensor sizes, even handling zooming in video (FMV drone video anyone?!). If you've done photogrammetry or played with NeRFs you know that is a HUGE deal. They've solved some tricky problems like efficient point matching and tracking, plus they've figured out how to fuse point clouds and handle loop closures in real-time. Their system runs at about 15 FPS on a 4090 and produces both camera poses and dense geometry. When they know the camera calibration, they get SOTA results across several benchmarks, but even without calibration, they still perform well. What's interesting is the approach -- most recent SLAM work has built on DROID-SLAM's architecture, but these folks went a different direction by leveraging a strong 3D reconstruction prior. Seems to give them more coherent geometry, which makes sense since that's what MASt3R was designed for. For anyone who cares about monocular SLAM and 3D reconstruction, this feels like a significant step toward plug-and-play dense SLAM without calibration headaches -- perfect for drones, robots, AR/VR -- the works!

Bilawal Sidhu

704,008 views • 1 year ago