Note: This is a dynamic event: our schedule evolves until and throughout the school.

Note: Classes will be streamed online through Microsoft Teams, see each session for the link. A recording of each session will be uploaded to the ACM Digital Library.

Day 1 | Monday, Aug 10

Time  
10:00 Arrival & Introduction
Kunwar Grover & Tobias Grosser
11:45 Coffee break
12:15 Compiler Introduction
Saday Sadayappan & Tobias Grosser
Slides
14:00 Lunch
16:00 MLIR IR Introduction
Théo Degioanni & Mathieu Fehr
Notebook 1
Notebook 2
ListLang compiler
List dialect interpreter
18:00 Break
19:00 Poster session (not streamed)
21:00 — End of day—

Day 2 | Tuesday, Aug 11

Time  
10:00 ODS
Markus Böck
Slides
11:45 Coffee Break
12:15 MLIR Fundamentals: Introduction to C++ IR Data Structures and APIs
Matthias Springer
Slides
14:00 Lunch
16:00 MLIR Fundamentals: Pattern-based Transformations
Matthias Springer
Slides
17:45 Coffee break
18:15 Core transformations
Markus Böck & Maximilian Bartel
Slides
20:00 — End of day—

Day 3 | Wednesday, Aug 12

Time  
10:00 MLIR Fundamentals: Understanding Interfaces
Matthias Springer & Kunwar Grover
Slides
11:45 Coffee Break
12:15 Mutable variables and Mem2Reg
Théo Degioanni
Slides
MemorySlotInterfaces.td
MemorySlotUtils.h
14:00 Lunch
17:00


21:00

Hike
&
Solar
Eclipse

22:00 — End of day—

Day 4 | Thursday, Aug 13

Time  
10:00 Dataflow Analysis
Mathieu Fehr
Slides
11:45 Coffee Break
12:15 Introduction to CUDA Tile IR
Théo Degioanni
Slides
14:00 Lunch
16:00 Digital Design with MLIR and CIRCT
Louis Ledoux
Slides
17:45 Coffee break
18:15 Linalg and Transform Dialect
Maximilian Bartel
Slides
20:00 — End of day—

Day 5 | Friday, Aug 14

Time  
  Participants’ presentations 1 (not streamed)
10:00-10:20 Alastair Reid – An MLIR backend for ISA specifications
10:20-10:40 Stefan Abi-Karam – AI-Driven Hardware Design With High-Level Synthesis (HLS) Compilers
10:40-11:00 Edward Erasmie-Jones – Towards a Domain-Specific Compiler for Computational Fluid Dynamics Using MLIR
11:00-11:20 Ioana Ghiban – On embedding ML inside LLVM: From TOSA, via EmitC, to LLVM
11:20-11:40 René Zander – Qrisp: From High-Level Quantum Programming to MLIR
11:40-12:00 Juan Gallego – MLIR-based compilation for automatic acceleration deployment on CGRAs
  Coffee Break
  Participants’ presentations 2 (not streamed)
12:20-12:40 Edward Stow – Compilation for Quantum Error Correction
12:40-13:00 Matt Arsenault – The Trouble With Triples
13:00-13:20 Aidan Dakhama – CONQuER: Hardware-Aware Mixed-Precision Quantisation with Online-Calibrated Surrogates
13:20-13:40 Björn Striebing – AI Microburst: From Cloud to the Far Edge in One Burst
13:40-14:00 Luca Parigi – An MLIR-based approach for lowering OpenMP constructs on heterogeneous runtimes
  Closing & Lunch

Speakers

Maximilian Bartel

M. Bartel

roofline

Markus Böck

M. Böck

ETH Zürich

Théo Degioanni

T. Degioanni

NVIDIA

Mathieu Fehr

M. Fehr

U. of Edinburgh

Tobias Grosser

T.Grosser

U. of Cambridge

Kunwar Grover

K. Grover

AMD

Louis Ledoux

L. Ledoux

INRIA

P.Sadayappan

P.Sadayappan

U. of Utah

Matthias Springer

M. Springer

NVIDIA