User Manual¶
How to write, compile, run, and debug PyPTO programs.
Contents¶
| Page | What it covers |
|---|---|
| Getting Started | Installation, a first tensor program, tile kernels, loops, multi-function programs, compiling and running, DeviceTensor and explicit dispatch |
| Language Guide | The type system, programs and functions, operations, SSA and control flow, memory and data movement, InCore scopes, compilation |
| Operation Reference | The operator surface across the pl.*, pl.tensor.*, and pl.tile.* namespaces |
| Torch Codegen Debug Guide | Generating a PyTorch reference implementation from the IR to isolate accuracy problems |
Reading path¶
- Getting Started — get something running end to end before reading anything else.
- Language Guide — the syntax and semantics behind what you just ran.
- Operation Reference — consult as needed while writing kernels.
- Torch Codegen Debug Guide — when the output does not match your reference.
What is not here yet¶
This manual is being expanded into a full chaptered structure — tutorials, distributed programming, performance optimization, and accuracy debugging each get their own chapter. Until those land, the corresponding material lives in the developer documentation:
| Topic | Current location |
|---|---|
Tasks, dependencies, manual_scope / submit |
Python IR Syntax Specification, AutoDeriveTaskDependencies |
| Distributed DSL and collectives | Distributed Operators |
| Performance hints and diagnostics | Diagnostics, Compile Profiling |
| Runtime DFX flags, ring sizing, memory map | Runtime DFX, Per-Task Ring Sizing, Memory Map |
| External C++ kernels | Integrating Hand-Written C++ Kernels |
See Also¶
- Developer documentation — how the compiler lowers what you write.
- PTO ISA reference — the instruction semantics behind the generated code.
- Runtime documentation — the scheduler that executes compiled programs.