Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

Digital Foundry discuss #Steam Machine specs, with concerns on performance and 8GB of VRAM. "The GPU is a little bit anemic perhaps, esp relative to current gen consoles, which it's trailing slightly" "8GB of VRAM poses more possibly fundamental limits"

NIB

6,631 subscribers

30,276 Aufrufe • vor 7 Monaten •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

Generate Songs with HeartMuLa, with 8GB VRAM! HeartMuLa is the latest open source song generation AI model everyone's raving about. But the default model couldn't run on low VRAM machines. Then WanGP just made it work with just 8GB VRAM! 1-Click Install on Pinokio.

Generate Songs with HeartMuLa, with 8GB VRAM! HeartMuLa is the latest open source song generation AI model everyone's raving about. But the default model couldn't run on low VRAM machines. Then WanGP just made it work with just 8GB VRAM! 1-Click Install on Pinokio.

cocktail peanut

27,848 Aufrufe • vor 4 Monaten

Anyone with 8GB or 12GB VRAM setups needs to understand that "-ncmoe" is the key flag to boost performance on llama.cpp Here are my results for Qwen3.6 35B A3B, with 64k q8_0 context on a 8GB RTX 3070Ti: ⚪️ no flag → 8.7 tok/s RAM: 13.6GB & VRAM: 7.8GB 🔴 -ncmoe 35 → 27.5 tok/s RAM: 12.1GB & VRAM: 4.3GB 🟢 -ncmoe 30 → 32.5 tok/s RAM: 12GB & VRAM: 5.6GB 🔵 -ncmoe 25 → 40.9 tok/s RAM: 12GB & VRAM: 6.9GB Please note the ram and vram usage you see are total usage of a windows pc, with the model running. My friend's setup: 8GB VRAM and 16GB RAM. You can boost performance by switching to Linux, just something to keep in mind. Basically, this flag keeps the MoE experts in the first X layers on your CPU + RAM, instead of eating all your VRAM straight away. This is a smart hybrid offload way that lets you run bigger models without OOM while keeping the rest on your GPU for speed. As we can see on the data, there's a sweet spot. When we lower it from 35 to 25, speed bumps +50% because there are more layers on your GPU (look at the VRAM usage). The key here is to play around with the number and fit as much as possible on your VRAM, goal is to have 1GB/800MB headroom to avoid stress. ↓ server flags below

Anyone with 8GB or 12GB VRAM setups needs to understand that "-ncmoe" is the key flag to boost performance on llama.cpp Here are my results for Qwen3.6 35B A3B, with 64k q8_0 context on a 8GB RTX 3070Ti: ⚪️ no flag → 8.7 tok/s RAM: 13.6GB & VRAM: 7.8GB 🔴 -ncmoe 35 → 27.5 tok/s RAM: 12.1GB & VRAM: 4.3GB 🟢 -ncmoe 30 → 32.5 tok/s RAM: 12GB & VRAM: 5.6GB 🔵 -ncmoe 25 → 40.9 tok/s RAM: 12GB & VRAM: 6.9GB Please note the ram and vram usage you see are total usage of a windows pc, with the model running. My friend's setup: 8GB VRAM and 16GB RAM. You can boost performance by switching to Linux, just something to keep in mind. Basically, this flag keeps the MoE experts in the first X layers on your CPU + RAM, instead of eating all your VRAM straight away. This is a smart hybrid offload way that lets you run bigger models without OOM while keeping the rest on your GPU for speed. As we can see on the data, there's a sweet spot. When we lower it from 35 to 25, speed bumps +50% because there are more layers on your GPU (look at the VRAM usage). The key here is to play around with the number and fit as much as possible on your VRAM, goal is to have 1GB/800MB headroom to avoid stress. ↓ server flags below

left curve dev

165,327 Aufrufe • vor 1 Monat

LTX 2.3 v1.1 on 8GB Tested on rtx 4060ti 16gb vram and 64gb ram workflow:

LTX 2.3 v1.1 on 8GB Tested on rtx 4060ti 16gb vram and 64gb ram workflow:

Stable Diffusion Tutorials

24,587 Aufrufe • vor 1 Monat

TripoSplat 1-Click Launcher for PC with 8GB VRAM [NVIDIA Only] I wrote a 1-click launcher for the official TripoSplat Gradio app, run it locally. - Super minimal codebase (near zero dependencies) - Completely open (MIT License), both open code and weights - And seems to need just around 8GB VRAM??

TripoSplat 1-Click Launcher for PC with 8GB VRAM [NVIDIA Only] I wrote a 1-click launcher for the official TripoSplat Gradio app, run it locally. - Super minimal codebase (near zero dependencies) - Completely open (MIT License), both open code and weights - And seems to need just around 8GB VRAM??

cocktail peanut

45,968 Aufrufe • vor 18 Tagen

Overworld Waypoint 1.5 Low VRAM 1 Click Launcher Found this gem yesterday at the top of Huggingface leaderboard. This recording is running on my Windows PC, realtime. Not sped up. 1X speed. Crazy. Try with 1-click launcher (Requires NVIDIA around 8GB VRAM)

Overworld Waypoint 1.5 Low VRAM 1 Click Launcher Found this gem yesterday at the top of Huggingface leaderboard. This recording is running on my Windows PC, realtime. Not sped up. 1X speed. Crazy. Try with 1-click launcher (Requires NVIDIA around 8GB VRAM)

cocktail peanut

12,394 Aufrufe • vor 1 Monat

A 1-billion-parameter AI real-time motion model that connects to a 1080P camera or uploaded videos, supports UE, Unity, and Blender, and requires 8GB of VRAM for real-time processing.

A 1-billion-parameter AI real-time motion model that connects to a 1080P camera or uploaded videos, supports UE, Unity, and Blender, and requires 8GB of VRAM for real-time processing.

CYANPUPPETS

35,483 Aufrufe • vor 4 Monaten

Moore's Law Is Dead CONFIRMs any video card with more than 8GB VRAM "END OF LIFE" due to deliberate constrained shortages as hardware is priorities for AI. 5060ti, 5070ti, 5080, 5090, basically all of them. He backs up Hardware Unboxed's earlier video who broke the story.

Moore's Law Is Dead CONFIRMs any video card with more than 8GB VRAM "END OF LIFE" due to deliberate constrained shortages as hardware is priorities for AI. 5060ti, 5070ti, 5080, 5090, basically all of them. He backs up Hardware Unboxed's earlier video who broke the story.

Grummz

763,903 Aufrufe • vor 5 Monaten

I know some bros want to see what happens when we push the value even further, so here we go: ⚪️ -ncmoe 25 → 41.8 tok/s RAM: 12GB & VRAM: 6.9GB 🔴 -ncmoe 23 → 43.8 tok/s RAM: 12.2GB & VRAM: 7.4GB 🟢 -ncmoe 21 → 38.6 tok/s RAM: 12.4GB & VRAM: 7.8GB 🔵 -ncmoe 19 → 19.8 tok/s RAM: 13.8GB & VRAM: 7.8GB As you can see, there's a sweet spot with the VRAM usage. Play around and monitor to find the right value for your setup, you can use llama.cpp web ui to monitor speeds easily Sweet spot seems to be 25-23 for 8GB VRAM ✅ +40tok/s for Qwen3.6 35B with 64k q8_0 context on a 8GB card is very impressive just by using base llama.cpp, and we didn't even try Turboquant, MTP or Dflash yet! I'll focus on these next 👀 (Server flags and setup in the quoted tweet)

I know some bros want to see what happens when we push the value even further, so here we go: ⚪️ -ncmoe 25 → 41.8 tok/s RAM: 12GB & VRAM: 6.9GB 🔴 -ncmoe 23 → 43.8 tok/s RAM: 12.2GB & VRAM: 7.4GB 🟢 -ncmoe 21 → 38.6 tok/s RAM: 12.4GB & VRAM: 7.8GB 🔵 -ncmoe 19 → 19.8 tok/s RAM: 13.8GB & VRAM: 7.8GB As you can see, there's a sweet spot with the VRAM usage. Play around and monitor to find the right value for your setup, you can use llama.cpp web ui to monitor speeds easily Sweet spot seems to be 25-23 for 8GB VRAM ✅ +40tok/s for Qwen3.6 35B with 64k q8_0 context on a 8GB card is very impressive just by using base llama.cpp, and we didn't even try Turboquant, MTP or Dflash yet! I'll focus on these next 👀 (Server flags and setup in the quoted tweet)

left curve dev

18,049 Aufrufe • vor 1 Monat

Free Blender AI motion capture plugin: Connects to 1080P camera, runs on a local computer with a 10-billion-parameter model, and requires 8GB VRAM for real-time processing. Supports both real-time capture and video uploads.

Free Blender AI motion capture plugin: Connects to 1080P camera, runs on a local computer with a 10-billion-parameter model, and requires 8GB VRAM for real-time processing. Supports both real-time capture and video uploads.

CYANPUPPETS

233,749 Aufrufe • vor 1 Jahr

Okay this is nuts. Reseller Baily ecom says they are selling 5090s and getting returns where the GPU core and VRAM is stolen... GPUs and VRAM are worth more than gold now.

Okay this is nuts. Reseller Baily ecom says they are selling 5090s and getting returns where the GPU core and VRAM is stolen... GPUs and VRAM are worth more than gold now.

Grummz

449,698 Aufrufe • vor 3 Monaten

A 1-billion-parameter motion model trained on NVIDIA CUDA, requiring 8GB of VRAM for real-time operation. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

A 1-billion-parameter motion model trained on NVIDIA CUDA, requiring 8GB of VRAM for real-time operation. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

CYANPUPPETS

22,931 Aufrufe • vor 7 Monaten

🚀 Major Update for ML-Sharp WebUI on Pinokio! 🌟 Native Mac Video Export (Zero CUDA) 🎥 10 New Camera Trajectories 💾 Low VRAM Mode (8GB GPU Support) ⚡️ Real-Time Previews Create stunning 3D gaussian splats from a single image.

🚀 Major Update for ML-Sharp WebUI on Pinokio! 🌟 Native Mac Video Export (Zero CUDA) 🎥 10 New Camera Trajectories 💾 Low VRAM Mode (8GB GPU Support) ⚡️ Real-Time Previews Create stunning 3D gaussian splats from a single image.

franzipol

22,749 Aufrufe • vor 5 Monaten

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

CYANPUPPETS

124,640 Aufrufe • vor 5 Monaten

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

CYANPUPPETS

18,317 Aufrufe • vor 6 Monaten

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

CYANPUPPETS

32,379 Aufrufe • vor 7 Monaten

A 1-billion-parameter motion model trained on NVIDIA CUDA, requiring 8GB of VRAM for real-time operation. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, enabling rapid creation of 3D animation data.

A 1-billion-parameter motion model trained on NVIDIA CUDA, requiring 8GB of VRAM for real-time operation. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, enabling rapid creation of 3D animation data.

CYANPUPPETS

103,279 Aufrufe • vor 7 Monaten

Blender AI Motion Capture Plugin — Connects to a 1080P webcam and runs locally on your computer. Powered by a 1-billion-parameter model, it requires an 8GB VRAM GPU for real-time processing. Supports both real-time capture and video upload. The full-featured version currently supports NVIDIA CUDA only, and requires DX11 or above.

Blender AI Motion Capture Plugin — Connects to a 1080P webcam and runs locally on your computer. Powered by a 1-billion-parameter model, it requires an 8GB VRAM GPU for real-time processing. Supports both real-time capture and video upload. The full-featured version currently supports NVIDIA CUDA only, and requires DX11 or above.

CYANPUPPETS

58,611 Aufrufe • vor 1 Jahr

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

CYANPUPPETS

28,846 Aufrufe • vor 5 Monaten

AMD's Ray Regeneration in Crimson Desert really tanks the performance. GPU RX 9070 XT, Ryzen 9800X3D and 32GB VRAM

AMD's Ray Regeneration in Crimson Desert really tanks the performance. GPU RX 9070 XT, Ryzen 9800X3D and 32GB VRAM

NikTek

206,263 Aufrufe • vor 3 Monaten

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

CYANPUPPETS

28,871 Aufrufe • vor 5 Monaten