Lines Matching refs:table

38 a compact way. Each entry in the table takes up a fixed number of bytes.
41 table. Then they contain one or more fixed length `{key}`/`{value}` records of `{R}` bits sorted by
42 `{key}`.  The table data is easily memory mapped and each record can be randomly accessed by
78 the "same prefix" property. All ranges with a common prefix can be stored in one "suffix table".
86 Suffix table block IDs are calculated by taking the prefix of the S2 cell ID being sought and
87 applying a fixed offset. The block info and block data for the cell's suffix table can be accessed
93 used to obtain the `{block ID}` of the block used to store the suffix table. The `{block ID}` is
99 When the `{block ID}` block is non-zero length, the block is interpreted as a packed table
100 (described above) which stores the suffix table's entries.
104 only one record in the table.
106 For more information on suffix table storage, see the Suffix Table Blocks section below.
114 The header contains format information needed to address the suffix table blocks
130 It is important to keep the number of bits used in each suffix table entry to a
167 table, so this means that it can be stored once per table and only the (small) `{TZ IDs set ID}`
173 The suffix table block is a packed table with shared information and one or more entries.
179                           referenced by all table entries.
182 Each record in the suffix table logically holds entries consisting of:
187 As with any packed table, each record in the packed table has a fixed width of `{R}` bits. The first
192 the `{prefix}` for the table and the `{key}`.
201 When searching for an S2 cell ID, the prefix is used to locate the correct suffix table. The suffix
202 bits from the S2 cell ID can be extracted. Since all data in the table is held at a single S2 level,
203 the suffix bits can be used to binary search the suffix table entries by looking for an entry