Term · Advanced
Tensor Cores
Specialized GPU units on Nvidia RTX cards that accelerate AI and machine-learning math.
Tensor cores are dedicated blocks inside Nvidia RTX GPUs built for the matrix math behind neural networks. They speed up AI features creators use daily: denoising, upscaling, generative video, and voice tools. The more tensor cores and VRAM a card has, the smoother local AI workflows run.
Also called RT tensor cores, AI cores, tensor core
Related
- GPUA processor built for massively parallel work, powering effects, playback acceleration, and rendering.
- VRAMDedicated memory on the GPU that holds frames, textures, and effects data during rendering.
- OptiXNvidia's ray-tracing engine that uses RTX cores to speed up 3D renders and denoising.
- AI UpscalingUsing AI to boost image or video resolution by inventing plausible new detail.
- NVIDIA RTX 4090NVIDIA's flagship last-gen GPU with 24GB VRAM for heavy creative work.
- Generative AIAI that produces new content, images, video, audio, text, from patterns it learned.