Closed KuangjuX closed 10 months ago
When the size of the application's tensor is smaller than the given tile partition size, it will result in a generated thread block count of 0, in turn, leading to code generation and runtime errors.
This change try to fix this bug.
Resolved by #56
When the size of the application's tensor is smaller than the given tile partition size, it will result in a generated thread block count of 0, in turn, leading to code generation and runtime errors.