Glossary › AI Computing
AI Computing Glossary
Thinking of running AI on your own PC? These are the terms that decide what your machine can and can’t do.
VRAM
The GPU’s dedicated memory — the single most important spec for running AI models locally.
LLM (Large Language Model)
The AI models behind ChatGPT-style assistants — many can run privately on your own PC.
CUDA
NVIDIA’s computing platform that nearly all AI software is built on — why NVIDIA GPUs dominate AI.
TFLOPS
A GPU’s raw maths throughput — relevant for AI speed, but only comparable within one generation.
Guide: RTX GPU Comparison — VRAM & Power for Every Card
Find which RTX cards have 12GB or 16GB of VRAM, and how they rank on raw performance.
Questions about these terms
- What is the difference between VRAM and RAM?
- How long will 12GB of VRAM last?
- Is 32GB of VRAM overkill?
- Can I run an LLM locally without a GPU?
- How much RAM do I need to run an LLM locally?
- Can I run an LLM locally on Windows?
- Is CUDA the same as a GPU?
- Does CUDA only work on NVIDIA cards?
- What is the difference between CUDA and PyTorch?
See all common PC questions.
Compare VRAM across every card in our RTX GPU ranking, then see what these specs cost in our high-VRAM gaming builds from $750 — the same RTX cards that game also run local AI.
More glossaries: PC Hardware · Gaming · PC Repair · All terms