High Context window Question 155 of 223

What is PagedAttention and why does it matter for serving?

GenAI / LLM · Speak this in 60–90 seconds · Faridabad & Delhi NCR

PICTURE THIS: A SENTENCE BECOMES TOKENS

The model does not read letters like humans. It reads these pieces, then predicts the next one.

Simple meaning

PagedAttention, as in vLLM, stores the KV cache in non-contiguous pages like virtual memory.

WHY — Tokens instead of words?

LLMs use tokens (not full words) because it helps them:

Context window questions separate

people who only read docs from people who shipped.

Keep it short, concrete,

and tied to GenAI / LLM work.

Stable token IDs

Each piece maps to a number the network can learn.

Fits the model

Fixed pieces are what transformers expect as input.

STEPS — What happens with tokens?

Before the model can read a sentence, it goes through these steps:

1
PagedAttention, as in vLLM,
stores the KV cache in non-contiguous pages like virtual memory.
2
That cuts fragmentation and
enables continuous batching of many sequences.
3
Higher GPU utilization lowers
cost per token at a given latency.
4
Context mix
Attention looks at nearby tokens together.
5
Next token
The model scores what should come next.
6
Decode
IDs turn back into readable text.

EXAMPLE — See it in action

Let's see how a real sentence is tokenized (tokens may vary by model):

Input text

“That cuts fragmentation and enables continuous batching of many sequences.”

Tokenized output

Thatcutsfragmentationandenablescontinuous

Token IDs (example)

2987408337471632900

Note: Actual tokens and IDs depend on the tokenizer (e.g., GPT, Llama, etc.).

Key takeaway

PagedAttention, as in vLLM, stores the KV cache in non-contiguous pages like virtual memory. That cuts fragmentation and enables continuous batching of many sequences.