Device & Precision Hardware Profile

NVIDIA GPU
CUDA Version --
PyTorch Version --
Precision Mode fp16
Compute Capability --
Multiprocessors -- SMs
GPU Temperature --°C
Cooling Fan Speed --
Power Draw -- W
Core / Mem Clock -- MHz

Live VRAM Gauge & Peak Marker

Allocated: -- MB
Reserved: -- MB
Peak: -- MB
Capacity: -- MB

Active Session Cache Debugger (Persistent Store)

0 Active Sessions
Session ID Prompt Snippet Model Tokens Activation Size Created At
No cached sessions in memory store.

Adaptive VRAM Memory Topology Grid

🟦 Model Weights 🩵 Activation/KV Cache ⬛ Free Buffer

Layer-by-Layer Activation & Parameter VRAM Footprint

Layer Name Footprint Size (MB) VRAM Share Bar
Run prompt analysis to observe per-layer memory footprint.