IMPORTANT: To view this page as Markdown, append `.md` to the URL (e.g. /get-started.md). For the complete documentation index, see llms.txt.
Skip to main content
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).

Mojo struct

Struct_kv_cache_pcp_unpack_rows_paged

struct Struct_kv_cache_pcp_unpack_rows_paged

Registers the kv_cache.pcp_unpack_rows.paged graph op.

Writes a context-parallel group's all-gathered packed prefix rows into a cache outside any group, each at its global position. See nn.kv_cache_gather.kv_cache_unpack_prefix_rows. The .scaled variant also writes each row's scales.

Implemented traitsโ€‹

AnyType, Deinitable, Movable

Methodsโ€‹

executeโ€‹

static def execute[col_offset: Int, dcp_size: Int, dcp_interleave: Int](gathered: T, row_offsets: T, kv_blocks: T, page_stride: T, cache_lengths: T, kv_lookup_table: T, max_prompt_length: T, max_cache_length: T, layer_idx: UInt32, context: DeviceContext)

Was this page helpful?