正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Looking for a fast method to render your large-scale game components? Our lastest High-Perf Graphics paper allows to render terrains, oceans and even entire planets in realtime on any modern GPU using compute shaders. Fun fact: we use a Concurrent Binary Tree as a GPU memory manager.

Anis Benyoub

1,388 subscribers

50,765 次观看 • 2 年前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

10 条评论

Anis Benyoub 的头像

Anis Benyoub2 年前

Curious? Paper: Source code: Project page:

Anis Benyoub 的头像

Anis Benyoub2 年前

The less compressed video is here !

Anis Benyoub 的头像

Anis Benyoub2 年前

The HPG presentation is here:

VoR 的头像

VoR2 年前

@HPG_Conf @KennyPirman

LunCo.eth(🌍🚀🌖) sp/acc 的头像

LunCo.eth(🌍🚀🌖) sp/acc2 年前

@HPG_Conf So cool! At some point in time we would like to implement it for Godot to be able to render Moon in real-time to design space missions!

DP 的头像

DP2 年前

@HPG_Conf @SheriefFYI

Jean-David Génevaux (TFC) 的头像

Jean-David Génevaux (TFC)2 年前

@HPG_Conf La seule question qui vaille : mais c'est quoi cette bande son ? Grosse vibe Deus Ex...?

Anis Benyoub 的头像

Anis Benyoub2 年前

@HPG_Conf Et non, c’est de l’OST de Hotline Miami!

ThunderOwl - Sci-Fi Assets. Owlified. 🇱🇻🇺🇦🌞 的头像

ThunderOwl - Sci-Fi Assets. Owlified. 🇱🇻🇺🇦🌞2 年前

@HPG_Conf @GameDevMicah

James 的头像

James2 年前

@HPG_Conf Will there be a demo project made with Vulkan?

相关视频

As promised, more insights on the webgl raytracer. It is based on this article/paper by Tero Karras: It uses Morton code to build a BVH binary radix tree on the GPU in realtime. #webgl #bvh #raytracing

As promised, more insights on the webgl raytracer. It is based on this article/paper by Tero Karras: It uses Morton code to build a BVH binary radix tree on the GPU in realtime. #webgl #bvh #raytracing

Nicolas Riciotti

44,019 次观看 • 3 年前

We just open-sourced 3DREAL ! A new render-to-real IC-LoRA for LTX-2.3 from fal: turn any 3D / CG / game render into high 3d render, photorealistic, film-quality video, keeping your exact composition and camera move using the first frame.

We just open-sourced 3DREAL ! A new render-to-real IC-LoRA for LTX-2.3 from fal: turn any 3D / CG / game render into high 3d render, photorealistic, film-quality video, keeping your exact composition and camera move using the first frame.

fal

216,603 次观看 • 3 个月前

From AI imagination to GPU acceleration - this new tutorial shows how to build a full Astro Zombie Horde from a single image to final render.🧟‍♂️ Using AI tools on the front end and the The Render Network on the back, the entire pipeline comes together in hours - not days. Watch the full breakdown:

From AI imagination to GPU acceleration - this new tutorial shows how to build a full Astro Zombie Horde from a single image to final render.🧟‍♂️ Using AI tools on the front end and the The Render Network on the back, the entire pipeline comes together in hours - not days. Watch the full breakdown:

The Render Network

11,463 次观看 • 11 个月前

We got Kostia of Inferra on the stream yesterday talking about his experience renting GPUs to render videos for his film production company to eventually building Inferra while participating in the EasyA Kickstart hackathon Inferra is building the exchange for GPU compute, turning raw graphics-card power into something you can trade and route like any liquid market.

We got Kostia of Inferra on the stream yesterday talking about his experience renting GPUs to render videos for his film production company to eventually building Inferra while participating in the EasyA Kickstart hackathon Inferra is building the exchange for GPU compute, turning raw graphics-card power into something you can trade and route like any liquid market.

gg.xyz

18,738 次观看 • 3 个月前

Revolutionizing Access to #GPU Power - One Node at a Time.⚡️ Our 10th Utility is coming out on June 20.🚨 BIG GPU Nodes Platform isn’t just another name in the crowded GPU cloud services space, we’re a game-changer. Our platform is purpose-built to bridge the gap between server owners and those hungry for raw, high-performance computing power. 🖥️ We specialize in connecting suppliers from individual server owners to enterprise-scale GPU rig operators with consumers who demand serious horsepower for AI, machine learning, data science, and other compute-heavy applications. Whether you’re offering a sleek single-GPU setup or a massive mining-style array with multiple GPUs, BIG GPU Nodes is your launchpad. We make it simple to rent out underutilized GPU resources and just as easy for users to find exactly what they need — fast, affordable, and tailored to their workload.⛓️‍💥 This is more than cloud computing. This is decentralized GPU power for the future of innovation. ♾️ This product will generate revenue and we will give back to our loyal $CNCT holders. ✨ We’re proud to have MESSIER | M87 as our trusted partner and advisor, backing us with this vision. #M87 🤝 $CNCT From the very beginning, we promised you ONE BIG ECOSYSTEM, and a seamless space where everything you need comes together. Now, that vision is becoming reality. Millions of users will soon be connected, working, creating, and thriving within it. 🌍

Revolutionizing Access to #GPU Power - One Node at a Time.⚡️ Our 10th Utility is coming out on June 20.🚨 BIG GPU Nodes Platform isn’t just another name in the crowded GPU cloud services space, we’re a game-changer. Our platform is purpose-built to bridge the gap between server owners and those hungry for raw, high-performance computing power. 🖥️ We specialize in connecting suppliers from individual server owners to enterprise-scale GPU rig operators with consumers who demand serious horsepower for AI, machine learning, data science, and other compute-heavy applications. Whether you’re offering a sleek single-GPU setup or a massive mining-style array with multiple GPUs, BIG GPU Nodes is your launchpad. We make it simple to rent out underutilized GPU resources and just as easy for users to find exactly what they need — fast, affordable, and tailored to their workload.⛓️‍💥 This is more than cloud computing. This is decentralized GPU power for the future of innovation. ♾️ This product will generate revenue and we will give back to our loyal $CNCT holders. ✨ We’re proud to have MESSIER | M87 as our trusted partner and advisor, backing us with this vision. #M87 🤝 $CNCT From the very beginning, we promised you ONE BIG ECOSYSTEM, and a seamless space where everything you need comes together. Now, that vision is becoming reality. Millions of users will soon be connected, working, creating, and thriving within it. 🌍

Fixera AI

26,682 次观看 • 1 年前

Are you a #RobloxDev looking for NPCs to feature in your new game? Or a YouTube creator looking for characters to render in your video's thumbnail? Catalog Avatar Creator now allows you to export your outfits into Roblox Studio!!! Simply type "/export" in-game, and you will be prompted to save your current avatar into your Roblox Inventory as a package - which you can then insert into Roblox Studio from the Toolbox! #Roblox

Muneeb

115,500 次观看 • 3 年前

▶️ Solidus Ai Tech GPU Marketplace: First Look Video! 🌟 Introducing our GPU Marketplace preview! Access high-quality computing power economically from our cutting-edge HPC Data Center in Europe. Choose from a wide range of GPU options from global vendors with just a few clicks. 🔹Marketplace Offers: - Easy management of your resources - Choosing your GPU Instance - Customize your OS Template - Perpetual token burn using a portion of the profits on every transaction 🔸Benefits: - Part of AITECH Ecosystem - Economical and efficient - Streamlined processes - Reach a global audience - Maximize your returns - Deploy with a single click Solidus Ai Tech GPU Marketplace, ensures a swift, effortless, and cost-effective journey for both users and vendors. Our platform will provide seamless access to GPU power, real-time performance insights, and robust management tools. Join us to transform your computing experience.

▶️ Solidus Ai Tech GPU Marketplace: First Look Video! 🌟 Introducing our GPU Marketplace preview! Access high-quality computing power economically from our cutting-edge HPC Data Center in Europe. Choose from a wide range of GPU options from global vendors with just a few clicks. 🔹Marketplace Offers: - Easy management of your resources - Choosing your GPU Instance - Customize your OS Template - Perpetual token burn using a portion of the profits on every transaction 🔸Benefits: - Part of AITECH Ecosystem - Economical and efficient - Streamlined processes - Reach a global audience - Maximize your returns - Deploy with a single click Solidus Ai Tech GPU Marketplace, ensures a swift, effortless, and cost-effective journey for both users and vendors. Our platform will provide seamless access to GPU power, real-time performance insights, and robust management tools. Join us to transform your computing experience.

AITECH CLOUD NETWORK

82,395 次观看 • 2 年前

We are excited to share a deeper look into the GPU Sharing Platform that will power AI applications on the InfraX V4 DApp. In this demo we showcase a GPU running and completing an image generation task. Our backend successfully leverages this GPU (and CPU) in real time to complete the job, earning the user money, and powering the infraX platform. Here’s how it works 👉 when you connect to the InfraX DApp, our platform automatically detects the GPU on your machine and immediately checks it eligibility and where it can be used. From there, each user dictates exactly how and when they want to contribute their GPU power by setting the days their PC is most available (throughout the night, for example). Additional, each user will be able to pause their GPU's connection to infraX whenever they need to use it locally, resuming with a just a single click. Once activated, your GPU seamlessly plugs into the InfraX network and begins fulfilling compute jobs, earning YOU money. Go on, take the $INFRA pill..

We are excited to share a deeper look into the GPU Sharing Platform that will power AI applications on the InfraX V4 DApp. In this demo we showcase a GPU running and completing an image generation task. Our backend successfully leverages this GPU (and CPU) in real time to complete the job, earning the user money, and powering the infraX platform. Here’s how it works 👉 when you connect to the InfraX DApp, our platform automatically detects the GPU on your machine and immediately checks it eligibility and where it can be used. From there, each user dictates exactly how and when they want to contribute their GPU power by setting the days their PC is most available (throughout the night, for example). Additional, each user will be able to pause their GPU's connection to infraX whenever they need to use it locally, resuming with a just a single click. Once activated, your GPU seamlessly plugs into the InfraX network and begins fulfilling compute jobs, earning YOU money. Go on, take the $INFRA pill..

infraX | $INFRA

30,615 次观看 • 1 年前

Fast View Synthesis of Casual Videos paper page: Novel view synthesis from an in-the-wild video is difficult due to challenges like scene dynamics and lack of parallax. While existing methods have shown promising results with implicit neural radiance fields, they are slow to train and render. This paper revisits explicit video representations to synthesize high-quality novel views from a monocular video efficiently. We treat static and dynamic video content separately. Specifically, we build a global static scene model using an extended plane-based scene representation to synthesize temporally coherent novel video. Our plane-based scene representation is augmented with spherical harmonics and displacement maps to capture view-dependent effects and model non-planar complex surface geometry. We opt to represent the dynamic content as per-frame point clouds for efficiency. While such representations are inconsistency-prone, minor temporal inconsistencies are perceptually masked due to motion. We develop a method to quickly estimate such a hybrid video representation and render novel views in real time. Our experiments show that our method can render high-quality novel views from an in-the-wild video with comparable quality to state-of-the-art methods while being 100x faster in training and enabling real-time rendering.

Fast View Synthesis of Casual Videos paper page: Novel view synthesis from an in-the-wild video is difficult due to challenges like scene dynamics and lack of parallax. While existing methods have shown promising results with implicit neural radiance fields, they are slow to train and render. This paper revisits explicit video representations to synthesize high-quality novel views from a monocular video efficiently. We treat static and dynamic video content separately. Specifically, we build a global static scene model using an extended plane-based scene representation to synthesize temporally coherent novel video. Our plane-based scene representation is augmented with spherical harmonics and displacement maps to capture view-dependent effects and model non-planar complex surface geometry. We opt to represent the dynamic content as per-frame point clouds for efficiency. While such representations are inconsistency-prone, minor temporal inconsistencies are perceptually masked due to motion. We develop a method to quickly estimate such a hybrid video representation and render novel views in real time. Our experiments show that our method can render high-quality novel views from an in-the-wild video with comparable quality to state-of-the-art methods while being 100x faster in training and enabling real-time rendering.

AK

20,668 次观看 • 2 年前

PowerInfer - a high-speed inference engine for deploying LLMs locally. Just came across this super interesting project on speeding up inference. It's not MoE but it's a simple approach that exploits the high locality in LLM inference to design a GPU-CPU hybrid inference engine. Hot-activated neurons are preloaded onto the GPU for fast access, while cold-activated neurons (the majority) are computed on the CPU. This approach significantly reduces GPU memory demands and CPU-GPU data transfer. It achieves an average token generation rate of 13.20 tokens/s, with a peak of 29.08 tokens/s, across various LLMs on a single NVIDIA RTX 4090 GPU. It's on only 18% lower than that achieved by a top-tier server-grade A100 GPU. It also significantly outperforms llama.cpp by up to 11.69x while retaining model accuracy. There is a lot more innovation around inference that's coming fast. Really encouraged by the study on sparse computation to enhance the computational efficiency of LLMs. It's now possible to use PowerInfer with Llama 2 and Faclon 40B. Mistral-7B support is coming soon!

PowerInfer - a high-speed inference engine for deploying LLMs locally. Just came across this super interesting project on speeding up inference. It's not MoE but it's a simple approach that exploits the high locality in LLM inference to design a GPU-CPU hybrid inference engine. Hot-activated neurons are preloaded onto the GPU for fast access, while cold-activated neurons (the majority) are computed on the CPU. This approach significantly reduces GPU memory demands and CPU-GPU data transfer. It achieves an average token generation rate of 13.20 tokens/s, with a peak of 29.08 tokens/s, across various LLMs on a single NVIDIA RTX 4090 GPU. It's on only 18% lower than that achieved by a top-tier server-grade A100 GPU. It also significantly outperforms llama.cpp by up to 11.69x while retaining model accuracy. There is a lot more innovation around inference that's coming fast. Really encouraged by the study on sparse computation to enhance the computational efficiency of LLMs. It's now possible to use PowerInfer with Llama 2 and Faclon 40B. Mistral-7B support is coming soon!

elvis

261,756 次观看 • 2 年前

We’ve been saying that OctaRender is the rendering breakthrough every artist has been waiting for, but today, we MEASURE that. We don’t just build a feature-rich platform — we build something that performs. Our goal is for artists to get results as quickly as possible, without compromising on quality. We are continually improving, optimizing, and benchmarking against other platforms. Here's an example of a render comparison between The Render platform and OctaRender. You can check out the comparison in render speeds in the video. Render Info: - Scene: 100 frames at 1024 samples - Rendered on: $RENDER (left) using their Tier 2 (highest speed level) and on OctaRender (right) Notes: The render on OctaRender uses Cycles, while on Render, it uses Octane. Both engines are comparable in terms of rendering capabilities. OctaRender with Cycles maintains compositing nodes.

We’ve been saying that OctaRender is the rendering breakthrough every artist has been waiting for, but today, we MEASURE that. We don’t just build a feature-rich platform — we build something that performs. Our goal is for artists to get results as quickly as possible, without compromising on quality. We are continually improving, optimizing, and benchmarking against other platforms. Here's an example of a render comparison between The Render platform and OctaRender. You can check out the comparison in render speeds in the video. Render Info: - Scene: 100 frames at 1024 samples - Rendered on: $RENDER (left) using their Tier 2 (highest speed level) and on OctaRender (right) Notes: The render on OctaRender uses Cycles, while on Render, it uses Octane. Both engines are comparable in terms of rendering capabilities. OctaRender with Cycles maintains compositing nodes.

OctaSpace

16,475 次观看 • 2 年前

$15,000 for a Mac sounds insane. Until you realize what Apple is actually selling. 256GB of unified memory. Not a huge GPU farm. Not an NVIDIA rack. Not 4 different machines. Just one box that can keep multiple large LLMs in memory and run them locally. That changes the equation for local AI. NVIDIA wins when you need raw GPU compute. Apple gets interesting when the bottleneck is memory capacity. Because once your model doesn't fit into 24GB, 32GB or 48GB of VRAM, buying a faster GPU doesn't solve the problem. You need more memory. And that's where machines like the Mac Studio start looking less crazy. The interesting question isn't: “Is a $15K Mac faster than an NVIDIA GPU?” It isn't. The real question is: “How much are you willing to pay to make huge local models actually fit?” That’s where unified memory gets interesting.

$15,000 for a Mac sounds insane. Until you realize what Apple is actually selling. 256GB of unified memory. Not a huge GPU farm. Not an NVIDIA rack. Not 4 different machines. Just one box that can keep multiple large LLMs in memory and run them locally. That changes the equation for local AI. NVIDIA wins when you need raw GPU compute. Apple gets interesting when the bottleneck is memory capacity. Because once your model doesn't fit into 24GB, 32GB or 48GB of VRAM, buying a faster GPU doesn't solve the problem. You need more memory. And that's where machines like the Mac Studio start looking less crazy. The interesting question isn't: “Is a $15K Mac faster than an NVIDIA GPU?” It isn't. The real question is: “How much are you willing to pay to make huge local models actually fit?” That’s where unified memory gets interesting.

KyzoroX

19,536 次观看 • 1 个月前

🚨 YOUR GPU IS PROBABLY WASTING MORE THAN YOU THINK. vLLM is built to squeeze far more useful work out of your GPU when serving LLMs. Running an LLM at scale isn’t just about having a powerful GPU. The real problem is how efficiently you use its memory and compute. That’s where vLLM comes in. → High-throughput LLM inference and serving → PagedAttention for smarter KV-cache memory management → Continuous batching to keep GPUs busy → Prefix caching + chunked prefill → OpenAI-compatible API out of the box → Supports a huge range of modern LLM architectures → Quantization support for running models more efficiently And the crazy part? You can start an OpenAI-compatible inference server with: `vllm serve ` So your application can talk to your own model almost like it’s talking to OpenAI. The bigger idea: Don’t just buy more GPUs. Make the GPUs you already have work harder. That’s why vLLM has become such a major project in LLM inference. 🔥 #vLLM #AI #LLM #Inference #GPU #MachineLearning #AIInfrastructure #OpenSource #AIAgents #Developers

🚨 YOUR GPU IS PROBABLY WASTING MORE THAN YOU THINK. vLLM is built to squeeze far more useful work out of your GPU when serving LLMs. Running an LLM at scale isn’t just about having a powerful GPU. The real problem is how efficiently you use its memory and compute. That’s where vLLM comes in. → High-throughput LLM inference and serving → PagedAttention for smarter KV-cache memory management → Continuous batching to keep GPUs busy → Prefix caching + chunked prefill → OpenAI-compatible API out of the box → Supports a huge range of modern LLM architectures → Quantization support for running models more efficiently And the crazy part? You can start an OpenAI-compatible inference server with: `vllm serve ` So your application can talk to your own model almost like it’s talking to OpenAI. The bigger idea: Don’t just buy more GPUs. Make the GPUs you already have work harder. That’s why vLLM has become such a major project in LLM inference. 🔥 #vLLM #AI #LLM #Inference #GPU #MachineLearning #AIInfrastructure #OpenSource #AIAgents #Developers

Vikas gupta

14,138 次观看 • 20 天前

Etched is deploying two new technologies in chip design: low-voltage inference and cluster-scale memory. CEO Gavin Uberti says they'll make their chips much more power-efficient and way, way faster than today's leading GPUs. He breaks it down: "We looked at a lot of early research directions, and we realized the key things that models need are way more compute and way faster memory." "If you think about inference, there are two key parts: prefill and decode. For prefill, it's a compute-bound problem. You need to have more FLOPS, more operations per second on each of your chips." "On our GPU, the bottleneck's actually thermals. You can't really run a GPU at more than around 50% of what it could theoretically do, or it'll melt." "So we're using a new technology today called low-voltage inference to try to solve this problem. You bring the voltage of the chip down dramatically, which allows us to have way, way better efficiency in terms of how much power is drawn per unit of math, and thus fit way way more flops onto the chip..." "For decode, it's all about bandwidth. Not just bandwidth on a chip, but bandwidth across your cluster. That's why we have this technology we call cluster-scale memory. It reduces the amount of time it takes to communicate from one chip to another dramatically." "As a result we can go use all of our HBM, HBM bandwidth, SRAM, SRAM bandwidth, and our scale-up domain as a single coherent pool. And that means if you're a user, you can go get much faster tokens per second, while still keeping your costs low."

Etched is deploying two new technologies in chip design: low-voltage inference and cluster-scale memory. CEO Gavin Uberti says they'll make their chips much more power-efficient and way, way faster than today's leading GPUs. He breaks it down: "We looked at a lot of early research directions, and we realized the key things that models need are way more compute and way faster memory." "If you think about inference, there are two key parts: prefill and decode. For prefill, it's a compute-bound problem. You need to have more FLOPS, more operations per second on each of your chips." "On our GPU, the bottleneck's actually thermals. You can't really run a GPU at more than around 50% of what it could theoretically do, or it'll melt." "So we're using a new technology today called low-voltage inference to try to solve this problem. You bring the voltage of the chip down dramatically, which allows us to have way, way better efficiency in terms of how much power is drawn per unit of math, and thus fit way way more flops onto the chip..." "For decode, it's all about bandwidth. Not just bandwidth on a chip, but bandwidth across your cluster. That's why we have this technology we call cluster-scale memory. It reduces the amount of time it takes to communicate from one chip to another dramatically." "As a result we can go use all of our HBM, HBM bandwidth, SRAM, SRAM bandwidth, and our scale-up domain as a single coherent pool. And that means if you're a user, you can go get much faster tokens per second, while still keeping your costs low."

TBPN

20,404 次观看 • 3 个月前

Over 2,000 NVIDIA B300 GPUs are going live in a mega-cluster on Aethir's decentralized cloud very soon. ☁️ This is next-level infrastructure that unlocks access for the most ambitious AI workloads and use cases. 💻 Enterprise-grade GPU compute, ready to scale with your ambitions. ⚡️ Listen to Mark Rydon, our Co-Founder, explain the importance of Aethir’s B300 deployment.

Over 2,000 NVIDIA B300 GPUs are going live in a mega-cluster on Aethir's decentralized cloud very soon. ☁️ This is next-level infrastructure that unlocks access for the most ambitious AI workloads and use cases. 💻 Enterprise-grade GPU compute, ready to scale with your ambitions. ⚡️ Listen to Mark Rydon, our Co-Founder, explain the importance of Aethir’s B300 deployment.

Aethir

12,867 次观看 • 6 个月前

GPU Mining Rig Build 🔧 At Hash AI, our team has a wealth of experience in all things crypto mining. We have been assembling more GPU rigs, in house, to expand our fleet of AI-optimised miners. These rigs are powered by: ⚡️NVIDIA RTX A5000 Graphics Cards ⚡️CORSAIR RM1000x Power Supply ⚡️Biostar Mining Motherboard ⚡️Vengeance 32GB 6400Mhz DDR5 RAM ⚡️Intel Core I9 Processor ⚡️High Quality Braided Wiring By leveraging these high-performance components alongside our AI algorithm, we are able to mine the most profitable coins at any given time, with maximal efficiency.

GPU Mining Rig Build 🔧 At Hash AI, our team has a wealth of experience in all things crypto mining. We have been assembling more GPU rigs, in house, to expand our fleet of AI-optimised miners. These rigs are powered by: ⚡️NVIDIA RTX A5000 Graphics Cards ⚡️CORSAIR RM1000x Power Supply ⚡️Biostar Mining Motherboard ⚡️Vengeance 32GB 6400Mhz DDR5 RAM ⚡️Intel Core I9 Processor ⚡️High Quality Braided Wiring By leveraging these high-performance components alongside our AI algorithm, we are able to mine the most profitable coins at any given time, with maximal efficiency.

Hash AI

13,656 次观看 • 1 年前

GPU tradeoff series: A100 is not much more powerful than 4090 🫠 GPU Perf and Price: - 4090: 330 fp16 TFLOPs, $1,749 - A100 (80GB): 312 fp16 TFLOPs, $20,000 > A100 is 11.4X more pricy Training speed for GPT-2(124M) with llm.c: - 4090: 153K tokens/s - A100 (80GB): 195K tokens/s > A100 is only 1.3X faster (both trained using a single card, A100 llm.c training is shown in the video, 4090 video is in the quoted tweet) Conclusion: 4090 has a much better cost vs performance ratio Why: As in the H100 vs. 4090 comparison, the biggest difference between A100 and 4090 is their GPU memory size/bandwidth and cross-GPU communication bandwidth, which does not matter too much if your model can fit into a single 4090. Specs: 4090: - GPU memory size: 24GB - memory bandwidth: 1 TB/s - communication bandwidth: 64 GB/s A100: - GPU memory size: 80GB - memory bandwidth: 2 TB/s - communication bandwidth: 900 GB/s Nvidia killed off NVLink (a high-speed communication link that connects GPUs) on 4090. (Jensen Huang smiling face) If multiple 4090s could be interconnected via NVLink, their performance would be closer to datacenter-grade A100 GPUs, even for training larger models. Additionally, 4090 isn't allowed in datacenters, that's how Nvidia makes 💰💰💰

GPU tradeoff series: A100 is not much more powerful than 4090 🫠 GPU Perf and Price: - 4090: 330 fp16 TFLOPs, $1,749 - A100 (80GB): 312 fp16 TFLOPs, $20,000 > A100 is 11.4X more pricy Training speed for GPT-2(124M) with llm.c: - 4090: 153K tokens/s - A100 (80GB): 195K tokens/s > A100 is only 1.3X faster (both trained using a single card, A100 llm.c training is shown in the video, 4090 video is in the quoted tweet) Conclusion: 4090 has a much better cost vs performance ratio Why: As in the H100 vs. 4090 comparison, the biggest difference between A100 and 4090 is their GPU memory size/bandwidth and cross-GPU communication bandwidth, which does not matter too much if your model can fit into a single 4090. Specs: 4090: - GPU memory size: 24GB - memory bandwidth: 1 TB/s - communication bandwidth: 64 GB/s A100: - GPU memory size: 80GB - memory bandwidth: 2 TB/s - communication bandwidth: 900 GB/s Nvidia killed off NVLink (a high-speed communication link that connects GPUs) on 4090. (Jensen Huang smiling face) If multiple 4090s could be interconnected via NVLink, their performance would be closer to datacenter-grade A100 GPUs, even for training larger models. Additionally, 4090 isn't allowed in datacenters, that's how Nvidia makes 💰💰💰

Yuchen Jin

234,951 次观看 • 2 年前

Check out this cinematic FullDome 8K render created in Unreal Engine by Ivan Elizarov using Camera Render Studio and the Cinematic Earth asset by HuguesCG. The domed presentation format is intended for use cases like Planetariums, immersive theatres and more🌎 "In this render, I wanted to demonstrate how a detailed planetary environment can look when presented in the FullDome format. Instead of viewing the scene on a conventional flat screen, the image is designed to surround the audience across the dome surface, creating a greater sense of scale, depth, and immersion" - Ivan Elizarov

Check out this cinematic FullDome 8K render created in Unreal Engine by Ivan Elizarov using Camera Render Studio and the Cinematic Earth asset by HuguesCG. The domed presentation format is intended for use cases like Planetariums, immersive theatres and more🌎 "In this render, I wanted to demonstrate how a detailed planetary environment can look when presented in the FullDome format. Instead of viewing the scene on a conventional flat screen, the image is designed to surround the audience across the dome surface, creating a greater sense of scale, depth, and immersion" - Ivan Elizarov

Unreal Engine

20,841 次观看 • 2 个月前