Passes

Every transformation PyPTO runs over the IR, numbered to match its position in the default pipeline.

Pass documentation is numbered so that reading it front to back walks the compilation pipeline in execution order. 01–43 are pipeline passes; 91+ is reserved for passes that run at several positions and for infrastructure that is not a pipeline pass at all.

Framework

Page	What it covers
Pass, PassContext, PassPipeline, and PassManager	Organizing and executing passes with property tracking, instrumentation, and strategy-based pipelines

Default pipeline

Order	Pass	What it does
01	InlineFunctions	Splices FunctionType.Inline bodies into every call site
02	UnrollLoops	Expands ForKind::Unroll loops at compile time
03	CtrlFlowTransform	Rewrites break / continue into structured control flow
04	ConvertToSSA	Converts to SSA form with variable renaming, phi nodes, and iter_args
05	Simplify	Folds arithmetic, shape expressions, and scalar constant bindings
06	FlattenCallExpr	Flattens nested call expressions into three-address form
07	OutlineHierarchyScopes	Outlines Hierarchy scopes into functions carrying level / role metadata
08	OutlineIncoreScopes	Outlines InCore scopes into separate functions
09	OutlineClusterScopes	Outlines Cluster scopes into Group functions and standalone Spmd scopes into Spmd functions
10	ConvertTensorToTileOps	Converts tensor ops to tile ops in InCore functions, updating orchestration call sites
11	OptimizeOrchTensors	Eliminates redundant orchestration allocations and improves data flow
12	LowerCompositeOps	Decomposes composite tile / distributed ops into primitives
13	FlattenTileNdTo2D	Flattens 3D+ tile operations to 2D by merging all but the last dimension
14	LegalizeTileCast	Expands tile.cast pairs the ISA cannot emit as one instruction into the shortest native chain
15	AutoTileMatmulL0	Picks an L0 tile shape (m, n, k) from the backend's L0 capacities and tiles matmuls to it
16	CanonicalizeTileSlice	Lowers tile.slice into the canonical tile.extract form
17	InferTileMemorySpace	Infers the on-chip MemorySpace of every tile and inserts tile.move to legalize mismatches
18	ResolveBackendOpLayouts	Repairs backend-required tile layouts for elementwise ops
19	LowerAutoVectorSplit	Converts AUTO pl.split mixed InCore functions into the explicit split_aiv form
20	ExpandMixedKernel	Splits mixed InCore functions into separate AIC (Cube) and AIV (Vector) kernels
21	InjectGMPipeBuffer	Injects the __gm_pipe_buffer workspace for GM-routed cross-core pipes (Ascend910B)
22	SplitVectorKernel	Stamps split attributes and handles the no-split dual-AIV path
23	StampTfreeSplit	Copies each cross-core tpop's split and pipe id onto its matching tfree op
24	NormalizeReturnOrder	Reorders every InCore function's return tuple into the canonical order
25	SkewCrossCorePipeline	Software-pipelines mixed cube/vector loops so the two cores overlap
26	LowerPipelineLoops	Replicates pl.pipeline(N, stage=F) bodies F times to enable ping-pong buffering
27	CanonicalizeIOOrder	Reorders pipeline-body statements along the scalar → load → compute → store ladder
28	MaterializeTensorStrides	Fills in the packed canonical stride for every tensor view that carries none
29	InitMemRef	Initializes MemRefs and creates alloc operations with unallocated addresses
30	MaterializeSemanticAliases	Forces buffers that program semantics require to be one allocation (loop-carry, in-place)
31	MemoryReuse	Reuses buffers by lifetime analysis and removes redundant allocs
32	AllocateMemoryAddr	Assigns real addresses to existing alloc operations
33	FoldNoOpReshape	Folds tile.reshape calls that change neither physical shape nor allocation
34	FuseCreateAssembleToSlice	Fuses tensor.create + tensor.assemble into one tensor.slice view
35	DeriveCallDirections	Materializes wrapper ParamDirections, then derives a per-argument ArgDirection at every call
36	AutoDeriveTaskDependencies	Derives conservative task-to-task dependency edges
37	ExpandManualPhaseFence	Compresses profitable full-array TaskId dependencies in manual scopes
38	SynthesizeAllReduceSignals	Turns a host allreduce's optional signal into explicit internal signal IR
39	MaterializeCommDomainScopes	Assembles WindowBuffer and CommDomainScopeStmt wrappers in each host orchestration body
40	LowerHostTensorCollectives	Rewrites host-level tensor collectives into internal builtin chip dispatches
41	MaterializeDistTensorCtx	Materializes an explicit CommCtx parameter and argument per DistributedTensor
42	MaterializeRuntimeScopes	Inserts AUTO RuntimeScopeStmt nodes so orchestration codegen emits PTO2_SCOPE 1:1
43	ClassifyIterArgCarry	Classifies each orchestration ForStmt iter_arg as a trivial alias or a materialised rebind carry

Outside the default pipeline

Page	What it covers
Utility Passes	Normalization and cleanup passes that run at several pipeline positions
Diagnostics	The advisory channel for compile-time warnings and performance hints
IR Verifier	Pluggable property verifiers that validate IR correctness between passes

Shared material

Page	What it covers
Shared Pass Utilities	Reusable helpers in include/pypto/ir/transforms/utils/
Loop-Carried Compiler Dependency Compression	How loop-carried dependency edges are compressed

See Also

• IR — the representation these passes transform.
• Backend — how passes get per-architecture answers without branching on the backend.
• Code Generation — what runs once the pipeline is done.