PC TOPS / TFLOPS
Calculator

What it is

The PC TOPS Calculator estimates a system’s theoretical compute power across three precisions:

• FP64 — 64-bit floating point (scientific computing, simulations)
• FP32 — 32-bit floating point (gaming, rendering, classic ML)
• INT8 — 8-bit integers via Tensor Core / NPU (AI inference, neural networks)

It’s not a benchmark — it’s a theoretical-peak calculation based on the components’ official technical specs.

For AI workloads INT8 is the usual figure, use that as your reference value.

How to use it

Manual mode: enter the CPU and GPU values directly (cores, clock, SIMD, Tensor multiplier). Results update in real time.

Automatic lookup mode: type the component name in the search field and hit Search. The tool queries Claude AI to retrieve the technical specs and fill in the fields automatically.

The automatic lookup uses an internal database of the main CPUs and GPUs currently on the market, around 120 models in total.

Calculation structure

CPU FP32 = (P-core × clock × SIMD lanes × 2 × SIMD units) + (E-core × clock × SIMD lanes × 2)
GPU FP32 = shader cores × 2 × boost clock
GPU INT8 = GPU FP32 × Tensor multiplier
CPU INT8 = CPU FP32 × 4 (estimate)

The Tensor multiplier depends on the GPU generation: RTX 30xx = 4×, RTX 40xx = 8×, RTX 50xx = 16×.