tilelang.cuda.pipeline¶

Attributes¶

cuda_pipeline

Functions¶

allow_warp_specialized([pass_ctx, target])

module_has_tma(mod)

Check if any function in the module was lowered with TMA operations.

CUDAPassPipelineBodyPrologue(mod, target)

CUDAPassPipelineBody(mod, target)

Module Contents¶

tilelang.cuda.pipeline.allow_warp_specialized(pass_ctx=None, target=None)¶
Parameters:

  • pass_ctx (tilelang.transform.PassContext | None)

  • target (tvm.target.Target | None)

Return type:

bool

tilelang.cuda.pipeline.module_has_tma(mod)¶

Check if any function in the module was lowered with TMA operations.

This reads the tl.has_tma attribute set by LowerTileOp during LowerAndLegalize, which is the source of truth for whether TMA copies were actually generated.

Parameters:

mod (tvm.IRModule)

Return type:

bool

tilelang.cuda.pipeline.CUDAPassPipelineBodyPrologue(mod, target)¶
Parameters:

  • mod (tvm.IRModule)

  • target (tvm.target.Target)

Return type:

tvm.IRModule

tilelang.cuda.pipeline.CUDAPassPipelineBody(mod, target)¶
Parameters:

  • mod (tvm.IRModule)

  • target (tvm.target.Target)

Return type:

tvm.IRModule

tilelang.cuda.pipeline.cuda_pipeline¶