Cache decisions · RecallWeave

Choose a request sequence

One shared sequence, two histories

0 requests · capacity 2 · each card shows the selected prefix

FIFO

First in, first out

Oldest inserted → newest inserted

hits: 0
misses: 0

LRU

Least recently used

Least recently used → most recently used

hits: 0
misses: 0

More room, same complete sequence

These totals use every request, regardless of the step selected above. Each row starts empty.

Whole-sequence misses by capacity
Capacity	FIFO misses	LRU misses

Inspect the complete trace

Cache contents are shown after each request, in each policy's order. “None” means no eviction.

All requests; the selected step is highlighted
Step	Page	FIFO	Evicts	FIFO contents	LRU	Evicts	LRU contents

Read the rule before the result

A hit finds the requested page already present. A miss loads it; if all slots are occupied, one page must leave first.

FIFO keeps the insertion order on a hit. LRU refreshes the hit to most recent. The highlighted page is the request you just inspected.

Neither rule knows future requests. LRU's capacity guarantee compares LRU with itself at different capacities; it does not make LRU the winner on every workload.

Take the next step

Turn the trace into a lesson

Predict the next eviction, explain the capacity anomaly and distinguish a trace count from a timing claim.

Download the original course JSON.
Open RecallWeave and choose Bring your own lesson.
Review the preview, then explicitly start the lesson.

12 questions · 4 concepts