"test/vscode:/vscode.git/clone" did not exist on "4dd9f036d39c27b56d6f523c03eed8008b53f71d"
  • Lei Wang's avatar
    [TMA] Bugfix when a shared buffer is both issued with tma store and tma load (#857) · b9a51c43
    Lei Wang authored
    - Updated `init_desc_arg_map` to use `Var` as the key instead of `String` in `lower_hopper_intrin.cc`.
    - Enhanced `func_call_args` method in `TLCUDASourceWrapper` to accept additional parameters for better argument mapping.
    - Added assertions to ensure consistency between function parameters and arguments during kernel launches.
    - Modified `generate_tma_descriptor_args` to utilize a mapping of variable names for TMA descriptor initialization.
    b9a51c43
lower_hopper_intrin.cc 5.89 KB