For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Python class

BlockCount

BlockCount

class max.pipelines.kv_cache.BlockCount(free, total)

source

Bases: object

A point-in-time snapshot of a block pool’s occupancy.

Used for the device (G0) pools, whose blocks are the unit the manager actually allocates in. The connector’s external tiers report ByteCount instead.

Parameters:

• free (int)
• total (int)

free

free: int

source

free_pct

property free_pct: float

source

Returns the percentage of blocks not currently in use, in [0, 100].

0 when total is 0, rather than dividing by zero.

total

total: int

source

used

property used: int

source

Returns the number of blocks currently in use.

used_pct

property used_pct: float

source

Returns the percentage of blocks currently in use, in [0, 100].

0 when total is 0 (such as a tier with no configured capacity), rather than dividing by zero.