FLOPS
In plain wordsFloating-point operations per second: how many arithmetic steps a computer can do each second. It is the standard speed rating for supercomputers and AI chips.
A rate of computation: the number of floating-point operationsFLOP, n.: A single arithmetic step, such as multiplying or adding two numbers. Counting them measures how much computing went into an AI model, and laws use such counts to decide which models face extra rules. performed per second, usually with SI prefixes — tera (10^12), peta (10^15), exa (10^18). Vendors quote peak FLOPS for each number format, and the figure rises as precision falls: an H100H100, prop. n.: Nvidia's flagship AI chip of 2022–2024, so scarce and sought-after during the generative-AI boom that it became a unit of measure for AI computing power. offers about 67 teraFLOPS in 64-bit tensor arithmetic, about 990 in bfloat16bfloat16, n.: A 16-bit number format invented at Google for AI. It keeps the huge range of the standard 32-bit format but drops precision, which neural networks can usually tolerate. and about 1,980 in FP8FP8, abbr.: An 8-bit number format that uses a quarter of the memory of the standard 32-bit format, letting new AI chips train and run models roughly twice as fast as with 16 bits. (dense figures; 'sparse' figures double them by assuming structured sparsitysparsity, n.: The property of being mostly zeros or mostly switched off. AI systems exploit it by skipping the empty parts, so they need less memory and computation.). Sustained performance is lower, which is why practitioners track model FLOPs utilizationmodel FLOPs utilization, n.: The share of a computer's top calculating speed that actually goes into useful AI training work. Real systems often reach only around a third to a half. and rooflinesroofline model, n.: A simple chart showing the fastest a program can run on a chip: limited either by how quickly data arrives from memory or by the chip's top calculating speed.. Supercomputers are ranked on the TOP500 by 64-bit LINPACK FLOPS — as of June 2026 China's LineShine led at about 2.2 exaFLOPS, one of five systems at or above one exaFLOPS — whereas 'AI exaFLOPS' claims for rack-scale systemsrack-scale system, n.: An AI computer designed as one whole cabinet of chips, cooling and wiring, rather than separate servers, so all the chips inside can work together like one big chip. count 4- or 8-bit operations.
Origin
'Flops' and 'megaflops' were in use in supercomputing by the mid-1970s; some sources credit David Kuck with coining them in 1974. Milestones on 64-bit LINPACK are the first teraFLOPS machine (ASCI Red, 1997), petaFLOPS (Roadrunner, 2008) and exaFLOPS (Frontier, 2022).
Usage note
Not the plural of FLOPFLOP, n.: A single arithmetic step, such as multiplying or adding two numbers. Counting them measures how much computing went into an AI model, and laws use such counts to decide which models face extra rules.: 'FLOPS' is a rate, 'FLOPs' a count. Peak figures are comparable only at the same precision and sparsity assumptions.
Related
- FLOPFLOP, n.: A single arithmetic step, such as multiplying or adding two numbers. Counting them measures how much computing went into an AI model, and laws use such counts to decide which models face extra rules.
- model FLOPs utilizationmodel FLOPs utilization, n.: The share of a computer's top calculating speed that actually goes into useful AI training work. Real systems often reach only around a third to a half.
- roofline modelroofline model, n.: A simple chart showing the fastest a program can run on a chip: limited either by how quickly data arrives from memory or by the chip's top calculating speed.
- TOPSTOPS, abbr.: Trillions of operations per second: the speed rating usually given for AI chips in phones and laptops, often counting simple whole-number math rather than precise fractions.
- AI supercomputerAI supercomputer, n.: A huge network of AI chips — often tens or hundreds of thousands — wired together in a data center to train the largest AI models.
- floating-point formatfloating-point format, n.: A way of storing numbers in a computer like scientific notation — a few digits plus a power of two — trading accuracy for range. AI uses ever-smaller formats to compute faster.
- FLOPFLOP, n.: A single arithmetic step, such as multiplying or adding two numbers. Counting them measures how much computing went into an AI model, and laws use such counts to decide which models face extra rules.
- TOPSTOPS, abbr.: Trillions of operations per second: the speed rating usually given for AI chips in phones and laptops, often counting simple whole-number math rather than precise fractions.
Mentioned in
FLOPFLOP, n.: A single arithmetic step, such as multiplying or adding two numbers. Counting them measures how much computing went into an AI model, and laws use such counts to decide which models face extra rules., AI supercomputerAI supercomputer, n.: A huge network of AI chips — often tens or hundreds of thousands — wired together in a data center to train the largest AI models., GPU-hourGPU-hour, n.: One graphics chip working for one hour — the everyday unit for how much computing a job used and for renting AI chips by the hour., memory bandwidthmemory bandwidth, n.: How fast data can be moved between a chip's memory and its calculators. For chatbots, this often decides how quickly words appear, more than raw calculating speed., tokens per secondtokens per second, n.: A speed rating for AI text generation: how many word-pieces a model produces each second. Five to ten keeps up with a fast reader; hundreds feel instant., training computetraining compute, n.: The total number of arithmetic operations performed to train a model. It is the standard yardstick for the size of an AI training effort, and some laws use it to decide which models face extra rules., arithmetic intensityarithmetic intensity, n.: How much calculating a task does for each piece of data it fetches from memory. Tasks with little calculation per fetch are slowed by memory, not by the chip's math speed., memory-boundmemory-bound, adj.: Describes a computer task slowed by how fast data can be fetched from memory rather than by how fast the chip can calculate — like a fast cook kept waiting for deliveries., MLPerfMLPerf, prop. n.: A standard set of speed tests for AI hardware and software, where chip makers and cloud companies compare how fast their systems train and run AI models., model FLOPs utilizationmodel FLOPs utilization, n.: The share of a computer's top calculating speed that actually goes into useful AI training work. Real systems often reach only around a third to a half., roofline modelroofline model, n.: A simple chart showing the fastest a program can run on a chip: limited either by how quickly data arrives from memory or by the chip's top calculating speed., TOPSTOPS, abbr.: Trillions of operations per second: the speed rating usually given for AI chips in phones and laptops, often counting simple whole-number math rather than precise fractions.