Equation 8 · Part 3 · The Deployment Envelope: Small Models Where the Power Is Not

Symbol B

T_{\mathrm{tok}} \ge \frac{W}{B}

B

What this part means

the achievable memory bandwidth and I the operational intensity of the kernel.

Its job in the formula

B occurs below the fraction bar. The quantity above the bar is divided by this expression; zero is excluded as a denominator.

Full expressionSymbol BArticle meaning

Where the article explains it

Their bound is one line, and it is the single most useful line in this whole subject: $R_{\mathrm{attainable}}$ = $\min(R_{\mathrm{peak}}, B \cdot I)$ with $R_{\mathrm{peak}}$ the peak arithmetic rate, B the achievable memory bandwidth and I the operational intensity of the kernel.

The passage around this formula

…generation and prefill behaving as different regimes [ 3 ] . The bound that matters on a device follows immediately: if W bytes of weights and cache must cross the memory bus for each token and the sustained bandwidth is B , then the time per token cannot be less than $T_{\mathrm{tok}} \ge \frac{W}{B}$ . no matter how fast the arithmetic units are. As an arithmetic illustration with assumed values rather than a measurement of any product: a four-billion-parameter model at four bits per weight is roughly two gigabytes of weights,…

Read this part in the article →

Learn the underlying idea

A variable is a named place for a value. Its letter is a local label: x can mean position in one formula and a data point in another.

Open the illustrated variables: a letter stands for a value guide →

See this notation across published equations →

Sources cited in the surrounding passage

[3] Efficiently Scaling Transformer Inference ↗

These citations provide research context; check each source for the exact claim it supports.