Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Just fixes comments #1998

Merged
merged 2 commits into from
Sep 27, 2022
Merged
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 4 additions & 6 deletions torch/csrc/jit/codegen/cuda/lower_alias_memory.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -24,15 +24,13 @@ IterDomain* exactConcreteId(IterDomain* id) {
id, IdMappingMode::EXACT);
}

//! Checks that the current loop nest is not realizing a serial
//! broadcast so that each index of producer buffer will only
//! be visited once, which is the only case where aggressive
//! inner sharing is valid.
//!
//! Checks that the current loop nest is realizing a serial
//! broadcast so that each index of producer buffer can be visited
//! multiple times, in which case the aggressive is not valid.
bool isSerialBroadcastResolution(TensorView* producer, TensorView* consumer) {
//! Note: see issue #1785:
//! serial broadcast resolution doesn't only happen to
//! immediate producers of broadcast ops. We can also have
//! immediate outputs of broadcast ops. We can also have
//! example:
//! T1[I,B] = broadcast(T0[I]])
//! T3[I,I] = T1[I,B] + T2[I,I]
Expand Down