mlir-tensorrt-v0.4.4dev20251231

What's Changed

• Integrate LLVM@d6e2143b064e by @christopherbate in #754
• [kernel] Add vector.from_elements unroll patterns to LowerToNVVM by @christopherbate in #755
• [compiler] Improve consistency of how 'target' attribute on grouping ops is handled by @christopherbate in #756
• [compiler] Move StableHLO partitioning attribute handling to StablehloToPlan pass by @christopherbate in #757
• [Plan] Rename cluster ops to shorter mnemonics by @christopherbate in #758
• [StablehloExt] Refactor simplification patterns into separate files by @christopherbate in #761
• integrate internal changes by @christopherbate in #762
• [CI] upgrade release base image from rockylinux8 to rocklinux9 by @lanluo-nvidia in #763
• [executor] Add SROA support by @christopherbate in #765
• [mlir-tensorrt] Integrate internal changes by @christopherbate in #768
• NFC: Add instructions for using custom llvm-project by @christopherbate in #767
• integrate internal changes by @christopherbate in #770
• [executor] Add executor.ctpop operation by @christopherbate in #769
• [executor] Handle type mismatches in getoffset lowering by @christopherbate in #771
• [CI] added ubuntu 22.04 container by @lanluo-nvidia in #764
• [compiler] NFC: consolidate Utils libraries from compiler to common by @christopherbate in #772
• [compiler] Add phase-start and phase-end options to main pipeline by @christopherbate in #773
• [integrations/PJRT] Improve symbol visibility control and fix error message by @christopherbate in #774
• [CI] a few CI changes by @lanluo-nvidia in #760

Full Changelog: mlir-tensorrt-v0.4.4.dev202512190...mlir-tensorrt-v0.4.4

mlir-tensorrt-v0.4.4.dev202512190

Major Features

• Open-source the kernel generation backend
• Open-source the PJRT (XLA/JAX) integration
• Various optimizations targeting StableHLO while loops and raising to for-style loops
• Improve C++ host code generation pipeline; fill in some missing EmitC converters

Generated Release Notes

• Update CI workflows and devcontainer configs for CUDA 12.9 and Ubuntu 24.04 by @christopherbate in #723
• migrate internal changes by @christopherbate in #722
• migrate internal changes by @christopherbate in #724
• [CI] Adding Test Pipeline for PRs by @lanluo-nvidia in #725
• [mlir-tensorrt] Migrate internal change by @christopherbate in #726
• Add conversion of JAX shape_assertion custom calls to cf.assert by @christopherbate in #728
• Add Math dialect to EmitC conversion support by @christopherbate in #729
• Refactor: Move StableHLO matchers to dialect utils and reorganize namespaces by @christopherbate in #730
• Fix zero-rank tensor handling in shape functions by @christopherbate in #731
• [compiler] Fix missing test file by @christopherbate in #732
• [compiler] Add stablehlo.get_dimension_size conversion in stablehlo-to-linalg by @christopherbate in #733
• [compiler] Add cf.assert support to convert-host-to-emitc by @christopherbate in #734
• NFC: make minor updates to pass naming by @christopherbate in #735
• [tensorrt] Make new transpose/reshape elimination pass as v2 pass. by @christopherbate in #736
• [compiler] Add mtrt-scf-strength-reduction pass by @christopherbate in #737
• integrate internal changes by @christopherbate in #738
• integrate internal changes by @christopherbate in #739
• [mlir-tensorrt] Integrate Kernel backend by @christopherbate in #740
• [mlir-tensorrt] Migrate PJRT integration by @christopherbate in #741
• [CMake] Fix TRT CUDA version selection and add tests by @christopherbate in #745
• [CI] Add release wheel pipeline by @lanluo-nvidia in #742
• [mlir-tensorrt] NFC: Integrate internal changes by @christopherbate in #746
• [compiler] Add mtrt-scf-float-strength-reduce pass to preprocessing pipeline by @christopherbate in #747
• NFC: Fix formatting in MLIR-TRT README.md by @christopherbate in #749
• [integrations/PJRT] Implement executable metadata queries by @christopherbate in #748
• [integrations/PJRT] Implement additional PJRT methods for executable serialization/deserialization by @christopherbate in #750
• NFC: Update MLIR-TRT build instructions for PJRT integration wheels by @christopherbate in #751
• Upgrade GitHub Actions for Node 24 compatibility by @salmanmkc in #743
• integrate internal changes by @christopherbate in #752
• [mlir-tensorrt] Update LLVM patch to add support for PTX ISA v9.0 by @christopherbate in #753

Full Changelog: mlir-tensorrt-v0.4.2dev20251112...mlir-tensorrt-v0.4.4.dev202512190

mlir-tensorrt-v0.4.2dev20251112

NFC: Fix CICD build script

tripy-v0.1.4

What's Changed

• [mlir-tensorrt] Raise Normalization for Pytorch (torch-mlir) Layer norm. by @matthewfl in #680
• [mlir-tensorrt] Raise activations from their elementwise representation to tensorrt.activation Op by @matthewfl in #679
• Migrate internal changes by @christopherbate in #687
• Add tensor methods for cast, copy, and shape manipulation ops by @akhilg-nv in #688
• [TensorRT] Extract layer metadata from FusedLocation by @yizhuoz004 in #689
• Removes a workaround for #496 when no-op functions were not compiled correctly. by @pranavm-nvidia in #692
• Adds support for reflect mode in the pad operation by @pranavm-nvidia in #696
• Use FusedLocation to hold layer metadata by @yizhuoz004 in #690
• Updates version to 0.1.4 by @pranavm-nvidia in #697

New Contributors

• @matthewfl made their first contribution in #680

Full Changelog: tripy-v0.1.3...tripy-v0.1.4

tripy-v0.1.3

What's Changed

• Add Stable Diffusion demo by @akhilg-nv in #100
• Update packages for MLIR-TRT 0.1.43 by @yizhuoz004 in #683
• Support DimensionInputInfo in tp.compile by @yizhuoz004 in #618
• Removes outdated alignment warning by @pranavm-nvidia in #684
• Update Tripy v0.1.3 by @yizhuoz004 in #685

Full Changelog: mlir-tensorrt-v0.1.43...tripy-v0.1.3

mlir-tensorrt-v0.1.43

[mlir-tensorrt] Update `CONTRIBUTING.md` (#681)

This PR updates `CONTRIBUTING.md` by adding information about spinning
dev container inside and outside of VS Code.

tripy-v0.1.2

What's Changed

• Various improvements to SAMv2 demo by @pranavm-nvidia in #646
• Adds an API to set dimension names, updates SAMv2 to trigger MHA fusion by @pranavm-nvidia in #657

Full Changelog: mlir-tensorrt-v0.1.42...tripy-v0.1.2

mlir-tensorrt-v0.1.42

Adds package index links for MLIR-TRT 0.1.42

mlir-tensorrt-v0.1.41

Updates MLIR-TRT version to 0.1.41

tripy-v0.1.1

What's Changed

• Add tensorrt_engine to Executable API by @yizhuoz004 in #608
• Fixes various issues in L1 tests, improves reliability of SAMv2 downl… by @pranavm-nvidia in #610
• Switches serialized_tensorrt_engine to a property by @pranavm-nvidia in #612
• Add support for TensorRT Quickly Deployable Plugins by @samurdhikaru in #578
• Simplifies silu, updates executables to reject tensors that are not GPU constants by @pranavm-nvidia in #614
• Reorders guides to make them flow more logically by @pranavm-nvidia in #615
• Implements various features/quality of life improvements to enable YOLO example by @pranavm-nvidia in #616
• Updates compile to set names for weights in the module by @pranavm-nvidia in #621
• Adds a Module API to initialize parameters to dummy values by @pranavm-nvidia in #623
• Updates Executables to track the InputInfo they were compiled with by @pranavm-nvidia in #622
• Update layernorm to directly use TRT API by @akhilg-nv in #624
• Updates version to 0.1.1 by @pranavm-nvidia in #625

New Contributors

• @samurdhikaru made their first contribution in #578

Full Changelog: tripy-v0.1.0...tripy-v0.1.1