subgroup_id

Check navigator.gpu?.wgslLanguageFeatures?.has('subgroup_id') before choosing a shader that uses this language feature. It is available automatically when supported; requires subgroup_id; documents the dependency and rejects the shader on unsupported implementations. This language token is not a requiredFeatures entry for requestDevice(); any additional device requirements are described below.

The @builtin(subgroup_id) input identifies the current subgroup within its workgroup, and @builtin(num_subgroups) reports how many subgroups that workgroup contains. These are workgroup-local values, distinct from subgroup_invocation_id, which identifies a lane inside a subgroup.

A reduction can elect one invocation per subgroup to write a partial result into workgroup memory at an index given by subgroup_id. A later workgroup-level step can combine those partial results. Size scratch storage conservatively and synchronize accesses before other invocations consume the results.

The browser’s language token is only part of the requirement: the device must enable subgroups, and the shader must contain enable subgroups;. Do not assume subgroup IDs or lanes have a particular relationship to local_invocation_index. Without this language feature, use an algorithm that does not require explicit subgroup numbering, such as a workgroup-memory reduction.