Reference

The main path runs through Parts I–IV. The Reference holds material you reach for while reading:

NeedWhere
Look up a TIRx language feature**[[GPU/modern-gpu-programming-for-mlsys/reference/TIRx Language Reference
Compiler internals (the lowering pipeline)**[[GPU/modern-gpu-programming-for-mlsys/reference/Compiler Internals
Debug asynchronous GEMM/FA hangs, crashes, wrong results, or slowdowns**[[GPU/modern-gpu-programming-for-mlsys/reference/Debugging Warp-Specialized Kernels

For the complete tvm.tirx Python API, see the upstream TVM documentation.

The TIRx native level (Introduction to TIRx) and the tensor layout model (TIRx Layout API) are covered in Part II.


Previous · Book index · Next