Skip to main content
EDGE AI · BASICS

What Does TOPS Mean in AI?

TOPS stands for tera operations per second: the number of trillion simple operations an AI processor can perform each second. It is the headline figure used to compare AI accelerators such as NVIDIA Jetson modules and NPUs. It is usually quoted for 8-bit integer (INT8) maths, and sometimes with sparsity, which doubles the figure.

01

The Short Answer

TOPS is a useful first filter, but two devices with the same figure can perform very differently. Memory bandwidth, how well the software is optimised for the hardware, the precision the model runs at and the power mode the device is set to all change real-world speed. Benchmarking the actual model is the reliable test.

  • //
    Check the precision INT8 figures are higher than FP16 or FP32.
  • //
    Check for sparsity Sparse figures assume models pruned to suit the hardware.
  • //
    Check the power mode Many modules quote peak TOPS at their highest power setting.
  • //
    Benchmark your model Real frames per second matter more than the headline.
02

Why It Matters

Our Orin Nano vs Orin NX FAQ compares two popular modules, and our Jetson Super Mode FAQ explains the newer power modes. The fanless GPU computer case study chose between Orin Nano and Orin NX on TOPS and power.