trunk/e97ee94e32f26ccec33db4c1411bddc4cc9e1a31: Let fastAtomicAdd derive its bound from a packed accessor (#196155)
- PyTorch: 2149 events in the last 90 days
- PyTorch: 2120th Release in the last 90 days
- Previous: earlier the same day · ciflow/trunk/200538: [UPDATE] Update
What happened
fastAtomicAdd needs a bound on how far it may look for a pairing partner when packing a 16-bit atomic. Callers holding a raw pointer supply that themselves, and getting it wrong is silent (too large lets the pairing write past the allocation) -- which is why ~20 scatter sites that subscript a packed accessor still use the plain atomic rather than hand-deriving a bound. Adds an accessor overload that derives offset a…
Summary assembled by rule from the sources below