meta data for this page
Differences
This shows you the differences between two versions of the page.
— |
enterprise:nvidia:description:gpu [2024/10/31 17:03] (current) mchus created |
||
---|---|---|---|
Line 1: | Line 1: | ||
+ | ====== NVIDIA GPU Cheatsheet ====== | ||
+ | |||
+ | ^ | ||
+ | | GPU memory | ||
+ | | FP64 | TFLOPS | ||
+ | | FP64 Tensor Core | TFLOPS | ||
+ | | FP32 | TFLOPS | ||
+ | | TF32 Tensor Core | TFLOPS | ||
+ | | BFLOAT16 Tensor Core | TFLOPS | ||
+ | | FP16 Tensor Core | TFLOPS | ||
+ | | FP8 Tensor Core | TFLOPS | ||
+ | | INT8 Tensor Core | TFLOPS | ||
+ | | GPU memory bandwidth | ||
+ | | Decoders | ||
+ | | Max thermal design power (TDP) | W | 250 | 300 | 400 | 400 | 700 | 350 | 400 | 700 | 700 | | ||
+ | | Multi-Instance GPUs | ||
+ | | Form factor | ||
+ | | NVLink | ||
+ | | PCIe | | Gen4 | Gen4 | Gen4 | Gen4 | Gen5 | Gen5 | Gen5 | Gen5 | Gen5 | |