Loading transcript...
Click for full transcript
Chapters & Sections (220)
▼
15:32
Initial Power Supply Considerations
16:46
GPU Installation and Troubleshooting
18:25
Power Supply Design and Features
21:01
GPU Performance Testing with Blender
22:04
Benchmarks and Comparison with Previous Results
23:14
Power Supply Options and SMB Bus Connection
24:15
Discussion on Electrical Standards and Variations
25:17
LM Studio and Linux Compatibility
27:11
Microsoft's Hash Verification Process and Latency Issues
28:17
Incorrect benchmarking of DeepSec R1 in BF-16
28:48
Native Q8 and GPU Performance
29:18
Distilled Model and Large File Size
31:01
Discussion on Harbor Info 64 and Temperature Issues
32:08
Troubleshooting Windows Issues and Open CL Benchmark
34:54
Prediction on GPU Performance and Secondhand Prices
35:25
Discussion on ADA Secondhand Prices and Cyberpunk Benchmark
36:38
Troubleshooting Steam Issues and Running Cyberpunk
37:09
Temperature Monitoring and Red Launcher Issues
38:16
Error 43 Solution for Video Card and Audio Driver
39:18
Rateracing Ultra Preset and GPU Performance for LLM
47:30
AI Model Download Issues and Performance Concerns
48:41
Blender Scores, Competex, and Personal Plans
49:45
AMD GPU and Water Block Discussion
50:16
AI Knowledge Gaps, Research, and Learning
51:19
Future of AI Models and Hardware
51:51
Excitement and Potential Concerns
52:23
Coupon Codes and HBM Customization
53:26
Mounting Bracket and Adapter Discussion
55:02
Benchmarks and Comparison to H100
56:04
Comparison to ADA Generation and Abuntu Switching Issues
57:07
Nvidia Inferencing Microservices (NIMS) and Cloud Compute
59:14
Gen 5 PCIe Switches and DJX Station Review
01:00:15
RDMA and Switch Options
01:01:18
Comparison of RDMA Chipsets
01:02:19
GPU Options for AI Workloads
1:03:51
Initial Model Loading Issues
1:05:26
VRAM and GPU Memory Concerns
1:06:32
Windows Issues and Workarounds
1:09:05
Model Regeneration and Optimization
1:12:08
AI-Generated Code and Math PhD Insights
1:22:06
Question on Drive Storage and Geekbench Installation
1:23:28
Geekbench Score Analysis
1:24:55
Model Directory and Geekbench Run
1:25:35
Cyberpunk Benchmarking and Comparison
1:28:39
Path Tracing and Ray Tracing Discussion
1:30:18
Comparison of Path Tracing Results
1:31:51
Auto Resolution Scaling and Frame Generation
1:32:25
Frame Tearing and Capture Issues
1:32:56
Deep Learning and Context Window Optimization
1:33:31
Cinematic Performance and Frame Tearing
1:34:14
AI Experiments and Rerunning the Blender Benchmark
1:34:45
Blender Optimization and Linux Compatibility
1:35:15
Perfect Numbers and AI Reasoning
1:36:48
Model Training and Optimization for Perfect Numbers
1:37:19
Benchmarking Model Sophistication and Information Organization
1:37:51
GPU Discussion and VGPU on Linux
1:38:21
Blender Optimization and GPU Subdivision
1:39:23
GPU Comparison and Linux Testing
1:41:34
Parameter Model and Intent Filtering
1:42:07
Model-Based Automation with Home Assistant
1:43:12
Comparison to Cloud-Based Services and Ellen Studio
1:44:15
Node Red and Alternative LLMs for Automation
1:45:15
NVIDIA Container Toolkit and Traditional Programming
1:46:21
CUDA, Containers, and Scalability
1:46:48
NVIDIA's Plans for a Marketplace
1:47:51
GPU and System Information
01:49:26
LM Studio and Analytics Concerns
01:49:57
Model Connection to Home Assistant Evolution
01:50:28
Home Assistant and Node Red Development
01:51:03
eGPU Nerdery and PCIe Gen 5 Testing
01:51:41
Whisper AI Input and Text-to-Speech
01:52:13
Loading Model and Nvidia Power Connector
01:52:44
Falcon Northwest System and HGP Question
1:53:44
GPU and Model Configuration
1:54:18
Model Load Failure and Error Handling
1:54:52
Performance Optimization and Layer Addition
1:55:50
Testing and Windows Environment
1:58:12
Model Comparison and RAM Usage Discussion
1:59:20
Modded GPU Testing and Alternative Options
1:59:51
AI and GIMPs Project Discussion
02:01:23
Kubernetes and Model Performance Comparison
02:03:09
RTS 6000 and Consumer Options for AI Models
02:04:35
Model Configuration and Performance Optimization
02:07:22
RAM Configuration and Model Selection
02:08:30
GPU and KV Cache Discussion
02:12:55
Troubleshooting Driver Issues and Reboots
02:14:52
Power Connector and PSU Discussion
02:16:26
Initial Troubleshooting Steps
02:16:58
Proposed Solution for GPU Issues
02:19:04
Discussion of Nvidia Driver Version and Compatibility
02:20:16
Activation Failure and Kernel Building
02:24:46
Camera and Mic Setup Discussion
02:27:49
Nvidia Driver Installation and Configuration
02:30:20
Nuking DRO and Software Packaging
02:31:35
Fixing Power Cable Issues
02:32:07
Silverstone Case and Linux Distro
02:32:40
NVIDIA Windows Driver for ARM
02:33:11
Falcon Northwest Frag Box and SSD Issues
02:33:42
LLM on Modded GPU and VRAM Limitations
02:34:12
GPU Borrowing and PCIe Cable Issues
02:35:19
Uninstalling Software and Chat Break
2:36:59
Linux and GPU Configuration Options
2:38:29
RTX Pro 6000 Acquisition and Usage
2:39:47
GPU Issues and Solutions
2:41:19
Phase Change Memory and Optane Discussion
2:43:29
Windows OS Drive Options
2:45:30
Firmware and GPU Configuration
02:47:40
Hot Plugging Thunderbolt GPU with Ryzen Linux Desktop
02:48:11
Oculink and NVL Link Discussion
02:48:42
HDMI DP Handshake Process and Linus Torvalds
02:49:20
Nvidia Driver Booting and Oculink Issues
02:49:51
Thread Ripper Motherboard Recommendations and AMD CPUs
02:52:35
Potential for Engineers to Learn Engineering in Minecraft
02:53:06
Learning RTL or Verilog for Engineering
02:53:38
Crossover Point between Survival Sims and Reality
3:13:04
Initial CUDA Problem Discussion
3:15:24
Nvidia Installation and Driver Issues
3:17:16
Multiple Version Installation and Virtual Environment
3:18:52
GPU Performance and Headroom Discussion
3:21:06
PLC Automation and Smart Home Systems
3:23:12
Initial Setup and Testing
3:24:48
Water Break and Chat Interaction
3:26:27
LLM Model Loading and Testing
3:30:36
GPU Usage and VRAM Discussion
03:47:26
Nvidia Inferencing Microser and Software Enablement
03:47:56
Power Usage Curve and Efficiency Comparison
03:51:07
LLM Testing and Memory Limitations
03:52:38
Power Reduction and Performance Loss
03:55:53
Temperature Check and Benchmarking Results
3:58:08
Raw Teraflops and LLM Performance
3:58:47
Windows Machine Turning Off
3:59:29
RTX 6000 and DeepSeek Issues
4:00:01
Switching to DeepSeek R1 and LLM Model
4:01:21
Comparison of DeepSeek Performance on Linux and LM Studio
4:02:23
Discussion on Quantization Approaches and Unsloth's Work
4:03:24
Conclusion on DeepSeek Performance and Efficiency
4:04:26
Performance Comparison and Cloud Access
4:04:58
AI Spirit Journey and Token Processing
4:05:30
Teraflop Computation and VRAM Usage
4:06:01
YouTube Algorithm and Stream Limitations
04:08:10
Cyberpunk Run and Server Room Plans
04:08:40
Loan to Steve and AI Workloads
04:09:11
Llama.cpp Updates and Optimizations
04:10:13
Closing Remarks and OBS Stop
1h 20m
3h 13m⏭ 2
48m⏭ 1
1h 40m⏭ 3
31m⏭ 1
1h 22m
1h 16m⏭ 4
1h 50m⏭ 3
47m
18m
4h 6m⏭ 2
2h 15m⏭ 2