Skip to content
Tagged articles

VRAM Optimization

4 articles · Page 1 of 1
IT Services Circle
IT Services Circle
Nov 5, 2026 · Game Development

DirectX 12 Agility SDK 1.722: 66× Ray Tracing FPS Boost, 90% VRAM Cut

Microsoft's DirectX 12 Agility SDK 1.722 preview introduces Clustered Geometry and PTLAS, delivering up to 66× higher ray tracing frame rates and 90% VRAM reduction on supported GPUs, while a separate analysis explains why raising Task Manager priority cannot fix system freezes.

Agility SDKClustered GeometryDirectX 12
0 likes · 9 min read
DirectX 12 Agility SDK 1.722: 66× Ray Tracing FPS Boost, 90% VRAM Cut
Old Zhang's AI Learning
Old Zhang's AI Learning
Oct 26, 2026 · Artificial Intelligence

Nanbeige4.3-33B vs Qwen3.8-27B: Benchmark Claims Meet Hardware Reality

The author benchmarks Nanbeige4.3-33B against Qwen3.8-27B on an A100-80GB GPU, finding Nanbeige wins select benchmarks but fails to deliver its claimed 512K context on a single card due to high KV cache overhead from its Llama architecture, suffers from excessive chain-of-thought tokens that break coding tasks, and lacks vision capabilities, while Qwen's linear attention enables longer context and multimodal support.

Chain-of-ThoughtLLM benchmarkingLinear attention
0 likes · 12 min read
Nanbeige4.3-33B vs Qwen3.8-27B: Benchmark Claims Meet Hardware Reality
Baidu Intelligent Cloud Tech Hub
Baidu Intelligent Cloud Tech Hub
Oct 15, 2026 · Artificial Intelligence

TB-Scale LLM Inference: Zero-Redundancy Weight Deployment & 80% KV Cache Compression

Baidu Baige presents extreme VRAM optimization for TB-scale model inference: zero-redundancy weight deployment via DP Attn-sharding (releasing 16–30 GiB/card) and full-route CPP (50% throughput gain), plus KV Cache compression via LU-KV (80% reduction, 0.52% accuracy loss), Draft SWA, and PD handshake optimization (20% VRAM release).

KV Cache compressionLLM InferenceLU-KV
0 likes · 27 min read
TB-Scale LLM Inference: Zero-Redundancy Weight Deployment & 80% KV Cache Compression
Black & White Path
Black & White Path
Aug 2, 2026 · Artificial Intelligence

Running a 2.8‑Trillion‑Parameter K3 Model on 4 GB VRAM with AirLLM

AirLLM introduces layer‑wise inference and per‑expert streaming to decouple VRAM usage from model size, enabling the 2.8‑trillion‑parameter Kimi K3 LLM to run on a single consumer‑grade GPU while preserving full‑precision accuracy and offering security‑focused insights.

AirLLMKimi K3Layer-wise Inference
0 likes · 9 min read
Running a 2.8‑Trillion‑Parameter K3 Model on 4 GB VRAM with AirLLM