Talks
Tutorials
- MLIR Tutorial ; slides - recording ; Mehdi Amini and River Riddle @ 2020 Virtual LLVM Developers
Past editions:
- MLIR Tutorial ; slides ; Jacques Pienaar, Sana Damani @ MLIR4HPC@LCPC 2019
- Tutorial: Building a Compiler with MLIR ; ( slides and recording) ; Mehdi Amini, Alex Zinenko, Nicolas Vasilache @ EuroLLVM 2019
Tech talks
About MLIR and MLIR Components
How to Build your own MLIR Dialect ; Marius Brehler @ FOSDEM 2023
Compiler Support for Sparse Tensor Computations in MLIR ; Aart Bik @ 2021 LLVM Developers’ Meeting November 16-19, 2021.
MLIR: Scaling Compiler Infrastructure for Domain Specific Computation ; Alex Zinenko @ CGO 2021.
LLVM Compiler Infrastructure in HPC Workshop Keynote: MLIR: an Agile Infrastructure for Building a Compiler Ecosystem ( slides - recording) ; Mehdi Amini
CGO 2020 Keynote: MLIR Compiler Infrastructure ; Chris Lattner, Tatiana Shpeisman
IMPACT 2020 Keynote: Polyhedral Compilation Opportunities in MLIR ; Uday Bondhugula
MLIR: Multi-Level Intermediate Representation for Compiler Infrastructure ; slides ; Tatiana Shpeisman, Chris Lattner @ EuroLLVM 2019
MLIR Primer: A Compiler Infrastructure for the End of Moore’s Law slides ; Chris Lattner, Jacques Pienaar @ Compilers for Machine Learning (C4ML) workshop at CGO 2019
Using MLIR
MLIR: Accelerating TF with compilers ; recording ; Jacques Pienaar (filling in for Tatiana Shpeisman) @ TF Dev Summit 2020
MLIR in the TensorFlow ecosystem ; slides ; Jacques Pienaar @ Compilers for Machine Learning 2020
MLIR: Accelerating AI ; recording ; Chris Lattner, Tatiana Shpeisman @ O’Reilly TensorFlow World 2019
An MLIR Dialect for High-Level Optimization of Fortran ; recording ; Eric Schweitz @ 2019 LLVM Developers’ Meeting - Bay Area
Swift as syntactic sugar for MLIR ; recording ; Eugene Burmako , Alex Suhan @ Scale By the Bay 2019
Targeting Accelerators with MLIR.jl ; recording ; James Bradbury @ JuliaCon 2019
Inside TensorFlow: MLIR for TF developers ; recording ; Jacques Pienaar
Compiling Ruby with MLIR ; recording ; Alex Denisov @ LLVM Social Berlin #20
Open design meeting presentations
We host a weekly public meeting about MLIR and the ecosystem. If you’d like to discuss a particular topic or have questions, please add it to the agenda doc.
The meetings are announced on Discourse, subscribing to this category is the best way to stay informed. You can also register to this public calendar.
About MLIR and MLIR Components
- 2024-08-22: Upstream SPIR-V Conversion slides recording
- 2023-02-15: OpenMP GPU target offload slides - recording
- 2023-12-14: Vector dialect, reshape, and handling of unit dimensions slides - recording
- 2023-11-16: Targeting H100 with NVGPU and NVVM Dialects slides - recording
- 2023-11-02: A MPI Dialect Proposal slides - recording
- 2023-10-26: A Polynomial Dialect Proposal slides - recording
- 2023-09-28: Sharding Framework for Device Mesh slides - recording
- 2023-07-13: mlir-query: Tool to query MLIR IR dynamically slides - recording
- 2023-07-06: RFC on Pattern Matching for Library and Acceleration Instruction Rewriting slides - recording
- 2023-06-22: Targeting ARM SME from MLIR and SME Dialect slides part 1 - slides part 2 - recording
- 2023-05-25: Distinct Attributes: Modeling LLVM’s distinct metadata in MLIR slides - recording
- 2023-03-23: Nelli, a lightweight, Pythonic, frontend for MLIR slides - recording
- 2023-03-09: Integrating Convolution Optimization into MLIR to Improve Performance Beyond Im2Col+GEMM slides - recording
- 2023-02-23: MLIR Actions: Tracing and Debugging MLIR-based Compilers slides - recording
- 2023-02-09: Deep Dive on MLIR Internals, Operation&Attribute, towards Properties slides - recording
- 2023-01-26: Controllable Transformations in MLIR ; slides - recording
- 2022-11-03: Poison Semantics in MLIR ; slides - recording
- 2022-06-02: IRDL, a dialect to represent IR definitions ; slides - recording
- 2022-03-17: Interfaces and dialects for precise IR transformation control ; slides - recording
- 2022-01-27: PyTACO - An End-to-End Use Case for the Sparse Tensor Compiler ; slides - recording
- 2022-01-27: Anatomy of Linalg.generic ; slides - recording
- 2022-01-13: One-Shot Function Bufferization of Tensor Programs ; slides - recording
- 2021-12-02: End-to-end flow for ML Compilation; slides - recording
- 2021-11-04: PDLL: a Frontend for PDL ; slides - recording ;
- 2021-09-09: An ML-Driven Autoconfigurator for Sparse Tensor Kernels in MLIR ; slides - recording
- 2021-08-26: High Performance GPU Tensor CoreCode Generation for Matmul Using MLIR ; slides - recording
- 2021-07-22: MLIR data visualization using PassInstrumentation ; slides - recording
- 2021-07-15: From MHLO To Linalg in IREE ; slides - recording
- 2021-06-24: Typedefs in MLIR ; slides - recording
- 2021-06-17: Defining operations, types, attributes, and dialects at runtime ; slides - recording
- 2021-06-10: MLIR Language Server (Modern IDE features for
.mlir
files) ; slides - recording - 2021-05-06: Alibaba Group: Disc revisit & future discussion ; slides - recording
- 2021-04-22: EmitC: Generating C/C++ from MLIR ; slides - recording
- 2021-04-15: Pattern Descriptor Language ; slides - recording
- 2021-04-01: Discussion about MLIR Bindings (C API, Python Bindings, other languages) status ; slides - recording
- 2021-03-18: MLIR AMX Vector Dialect ; slides - recording
- 2021-03-11: Sparse Tensor Type Discussion ; recording
- 2021-02-18 Compiler Support for Sparse Tensor Computations: A Status Update ; slides - recording
- 2021-02-04 Discussion about Implicit Attribute Propagation ; recording
- 2021-01-21 Data Layout RFC ; recording
- 2021-01-14 Dialect Conversion & Type Conversion ; recording
- 2020-11-19: Type Conversions the Not-So-Hard Way: MLIR’s new composable bufferize passes ; slides - recording
- 2020-09-24: Buffer Allocation in MLIR ; slides - recording
- 2020-09-10: TF Kernel Generator ; slides - recording
- 2020-08-13: Structured and Retargetable Compilation With Vectors ; slides - recording
- 2020-07-30: SPIR-V dialect and lowerings ; slides - recording
- 2020-07-23: Asynchronous execution ; slides - recording
- 2020-07-09: MLIR Test Case Reducer Tool ; RFC Starting in-tree development of python bindings ; and RFC Rebooting C APIs for core IR - recording
- 2020-06-25: Basic GPU Compute Algorithm ; slides - recording
- 2020-04-16: MLIR on GPUs ; slides - recording
- 2020-04-09: Technical discussion on TCP ; slides - recording
- 2020-03-05: Shape Dialect ; slides - recording
- 2020-01-24: Thoughts on Tensor Code Generation in MLIR ; slides - recording
- 2019-12-19: Interpreted Pattern Match Execution ; slides - recording
- 2019-12-12: GPUs in MLIR ; slides - recording
- 2019-12-05: Structured Ops ; slides - recording
- 2019-10-10: Interfaces in MLIR ; slides - recording
- 2019-10-03: Layout changes to MLIR Tensor and MemRef Types ; slides - recording
- 2019-08-29 & 2019-09-12: Table-driven Op Definitions and Rewrites ; slides part 1 / part 2 - recording part 1 / part 2
- 2019-08-08: What is an MLIR Dialect? ; slides - recording
- 2019-07-25: MLIR Operation Legalization Framework ; slides - recording
- 2019-07-18: A proposal to add layout to Tensor ; slides - recording
Using MLIR
- 2024-09-12: XLA Indexing Maps slides recording
- 2023-12-22: PyDSL, a small subset of Python for constructing affine & transform dialects slides recording
- 2023-06-08/15: Update on Numba/MLIR slides recording
- 2023-05-04: Catalyst, an AOT/JIT compiler for hybrid quantum programs slides recording
- 2023-03-30: Representing dataflow with MLIR ; slides - recording
- 2022-12-15: Embedded reactive programming in MLIR ; slides - recording
- 2022-10-06: Coordinate Transformations in AMD’s rocMLIR ; slides - recording
- 2022-08-11: HeteroCL, an MLIR-based Intermediate Representation for Accelerator Design with Decoupled Customizations ; slides - recording
- 2022-02-17: IREE’s ML input dialects ; slides - recording
- 2022-01-20: A Data-Centric Model for Performance Portability on Heterogeneous Architectures ; slides - recording
- 2021-10-7: The Torch MLIR Project ; slides - recording
- 2021-05-27: Quantum-Classical Compilation with MLIR ; slides - recording
- 2021-04-29: Tensor Processing Primitives ; slides - recording
- 2021-03-04: MLIR based Numba backend ; slides - recording
- 2021-02-11: Polygeist: An affine C frontend for MLIR ; slides - recording
- 2020-11-05: COMET: Domain Specific Compilation for Computational Chemistry ; slides - recording
- 2020-10-29: Using MLIR for Multi-Dimensional Homomorphism ; slides - recording
- 2020-10-22: Fusion of Elementwise operations with Linalg on tensors ; slides - recording
- 2020-10-01: NPComp update ; slides - recording
- 2020-08-20: IREE CodeGen ; slides - recording - transcript
- 2020-08-06: MultiLevel Tactics: Lifting loops in MLIR ; slides - recording
- 2020-05-21: MLIR case study: AVX512 codegen for Vector dialect ; slides - recording
- 2020-05-14: RISE: A Functional Pattern-based Dialect in MLIR ; slides - recording
- 2020-04-23: LLHD: A Multi-level Intermediate Representation for Hardware Description ; slides - additional slides - recording ;
- 2020-03-26: Tile Dialect ; slides - recording
- 2020-03-19: TFRT Deep Dive ; slides - recording
- 2020-02-27: End-to-end Dynamic Shape Support in MLIR ; slides - recording
- 2020-01-30: IREE: MLIR-based End-to-End ML Tooling ; slides - recording
- 2020-01-16: nGraph Dialect: High-Level Graph Optimizations for Deep Learning Workloads ; slides - recording
- 2019-11-21: TensorFlow Saved Model Dialect ; slides - recording
- 2019-11-14: A Compiler Intermediate Representation for Stencils ; slides - recording
- 2019-11-07: The Stripe dialect: An alternate approach for expressing affine computations ; slides - recording
- 2019-09-19: S4TF (Swift for TensorFlow): Swift as syntactic sugar for MLIR ; slides - recording
- 2019-08-22: Tensor Attribute Layout in MKL DNN ; slides - recording
- 2019-08-15: Building TensorFlow Converter Tools with MLIR ; slides - recording
- 2019-08-01: TensorFlow Graph Representation in MLIR ; slides - recording
Upcoming talks or presentations
The schedule for the open design meetings and upcoming presentations can be found in the agenda doc.
Past conferences and workshops
- 2020/02/22: C4ML 2020: Compilers for Machine Learning workshop co-located with CGO in San Diego (Multiple MLIR talks!)
- 2020/01/20: AccML 2020: HiPEAC Workshop on Accelerated Machine Learning.