mirror of
https://github.com/ml-explore/mlx.git
synced 2025-12-16 01:49:05 +08:00
Custom cuda kernel (#2517)
This commit is contained in:
committed by
GitHub
parent
f4c8888cbe
commit
e397177f6e
@@ -267,7 +267,8 @@ void Compiled::eval_gpu(
|
||||
}
|
||||
}
|
||||
|
||||
return std::make_pair(std::move(builder.os), std::move(kernel_names));
|
||||
return std::make_tuple(
|
||||
false, std::move(builder.os), std::move(kernel_names));
|
||||
});
|
||||
|
||||
// Collapse contiguous dims to route to a faster kernel if possible. Also
|
||||
|
||||
Reference in New Issue
Block a user