Consensus for TT: Buckets with Entries, or Entries with Buckets?

[mvanthoor]

Hi,

This may be an insanely stupid question, because it has no bearing on how a transposition table works; it's mainly a naming issue. What is the consensus about entries and buckets?

1. Does a TT have Entries, and each Entry has multiple Buckets?

OR

2. Does a TT have Buckets, and each Bucket has multiple Entries?

Even the CPW is unclear:

"This family of strategies is similar to the two-tier system, but any number of tiers (known as "buckets") can be used (typically the number is based on the size of a cacheline). The difference is that the buckets are not specific to one consideration, but rather the new entry overwrites the entry in the bucket with the lowest depth [23] . "

- A TT can have "tiers", which are known as "buckets" (which implies that a bucket is inside an entry, and thus an entry has mulitple buckets).
- The new entry overwrite the entry in the bucket (which implies that a bucket has multiple entries)

edit: the current dev-version has Entries in the TT containing multiple Buckets. Alpha 3 had Buckets, containing mulitple Entries per Bucket, so I swapped this around from option 2 to 1.

[hgm]

Usual terminology is that an entry contains info about a single position. A bucket is a group of slots to which a particular position could map in the hashing process. Strictly speaking a slot is a chunk of memory, and an entry the information stored in the slot, but these terms are often used as synonyms.