Course Review Slides Exams Sources

Sources

PL-tpuWriting TPU kernels
PL-gridGrids and BlockSpecs
PL-pipeTPU pipelining
PL-mmTPU matmul tutorial
PL-sparseScalar prefetch and block-sparse
PL-distDistributed Pallas
PL-coreCore-specific programming
PL-gpuPallas Mosaic GPU
K-tokamaxTokamax kernel library
SB9How to Profile TPU ProgramsThe JAX profiler
SB9How to Profile TPU ProgramsReading an XLA op
SB9How to Profile TPU ProgramsExample profile
SB9How to Profile TPU ProgramsProblems
XP-homeXProf
XP-traceTrace Viewer
XP-opprofHLO Op Profile
XP-roofRoofline Model
XP-memMemory Viewer
XP-jaxProfiling JAX programs
R-tpuinf-proftpu-inference profiling guide
X-tiledTiled layout
X-asyncAsync HLO ops
A-nkiNKI programming model
SB1All About RooflinesTile intensity

Chapter 10. Day 5, second block, 55 minutes

Pallas, XProf and a design review

Read Quiz12 Flashcards21 Teach-back7

21 left in this session, 21 of them new.Reviewed 0

Question

What is Pallas, and how does it reach a TPU?

TPU Deep Dive

A free course on TPU hardware and software for GPU engineers.

Course

Course map Slides Exams Review flashcards Moving off PyTorch/XLA

About

About Sources Privacy policy Terms of use

Project

Source on GitHub Report an error Your data

An independent educational project, not affiliated with or endorsed by Google, NVIDIA or AWS. Product names are trademarks of their owners. Sources accessed October 2026; figures marked provisional, derived or unverified are explained where they appear. Written with AI assistance and may contain errors: report one on GitHub.

© 2026 learntpu.dev. Content CC BY 4.0, code MIT.