v2.15.0

List view

Implemented executorch's multi-optimization profile

cla signed

component: api [C++]Issues re: C++ APIIssues re: C++ API

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: runtime

component: testsIssues re: TestsIssues re: Tests

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open (in progress).

pytorch/TensorRTnumber 4441

#4441 In pytorch/TensorRT;

· cehongwang opened on Jul 28, 2026

chore: add moe tests

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Force All Tests[L0+L1+L2]For run all the L0, L1, L2 testsFor run all the L0, L1, L2 tests

Status: Open (in progress).

pytorch/TensorRTnumber 4262

#4262 In pytorch/TensorRT;

· zewenli98 opened on May 13, 2026

feat: support Global Performance Tuning

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open (in progress).

pytorch/TensorRTnumber 4476

#4476 In pytorch/TensorRT;

· zewenli98 opened on Aug 11, 2026

Merged main

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4501

#4501 In pytorch/TensorRT;

· micwill755 opened on Aug 17, 2026

Add one-shot export+decomp path for dynamo compile.

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

Status: Open (in progress).

pytorch/TensorRTnumber 4474

#4474 In pytorch/TensorRT;

· micwill755 opened on Aug 10, 2026

[MD-TRT] Llama export TP example with KV cache

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: runtime

Status: Open (in progress).

pytorch/TensorRTnumber 4287

#4287 In pytorch/TensorRT;

· apbose opened on May 21, 2026

remove downcast of int64 to int32

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: convertersIssues re: Specific op convertersIssues re: Specific op converters

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Status: Draft (not ready).

pytorch/TensorRTnumber 4023

#4023 In pytorch/TensorRT;

· apbose opened on Jan 21, 2026

feat: add cutile_op for AOT QDP plugins from cuda tile kernels

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: build systemIssues re: Build systemIssues re: Build system

component: testsIssues re: TestsIssues re: Tests

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open (in progress).

pytorch/TensorRTnumber 4463

#4463 In pytorch/TensorRT;

· bowang007 opened on Aug 6, 2026

fix: optimize attention decomposition

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: convertersIssues re: Specific op convertersIssues re: Specific op converters

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4448

#4448 In pytorch/TensorRT;

· fs-eire opened on Jul 30, 2026

fix: skip const-fold for arange for attention mask

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4452

#4452 In pytorch/TensorRT;

· fs-eire opened on Jul 30, 2026

Cache post-lowering ATen graphs to skip decomp on warm compiles.

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Status: Draft (not ready).

pytorch/TensorRTnumber 4540

#4540 In pytorch/TensorRT;

· micwill755 opened on Aug 20, 2026

dryrun stats observable

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4519

#4519 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

associative_scan lowering

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4518

#4518 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

MLA attention fallback

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4517

#4517 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

MoE histc / _grouped_mm

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: convertersIssues re: Specific op convertersIssues re: Specific op converters

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4514

#4514 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

Fix strided slice_scatter KV cache updates

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: convertersIssues re: Specific op convertersIssues re: Specific op converters

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4550

#4550 In pytorch/TensorRT;

· apbose opened on Aug 22, 2026

complex _reshape_copy / _to_copy

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4516

#4516 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

Fix complex rewrite device mismatch

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4513

#4513 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

Fix CPU-only import

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

Status: Open (in progress).

pytorch/TensorRTnumber 4504

#4504 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

Fix noncontiguous AOT inputs

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4509

#4509 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

Fix cumsum dtype

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: convertersIssues re: Specific op convertersIssues re: Specific op converters

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4511

#4511 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

Fix TopK rank-1

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: convertersIssues re: Specific op convertersIssues re: Specific op converters

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4508

#4508 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

Fix slice_scatter open end

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: conversionIssues re: Conversion stageIssues re: Conversion stage

component: convertersIssues re: Specific op convertersIssues re: Specific op converters

component: coreIssues re: The core compilerIssues re: The core compiler

component: dynamoIssues relating to the `torch.compile` or `torch._dynamo.export` pathsIssues relating to the `torch.compile` or `torch._dynamo.export` paths

component: loweringIssues re: The lowering / preprocessing passesIssues re: The lowering / preprocessing passes

component: testsIssues re: TestsIssues re: Tests

Status: Open (in progress).

pytorch/TensorRTnumber 4512

#4512 In pytorch/TensorRT;

· jloftin-nv opened on Aug 18, 2026

Track the ExecuTorch pin and move it to new nightlies automatically

ci: nightlyRun the nightly lane (all tiers incl. llm / kernels / distributed) on every pushRun the nightly lane (all tiers incl. llm / kernels / distributed) on every push

cla signed

component: api [Python]Issues re: Python APIIssues re: Python API

component: build systemIssues re: Build systemIssues re: Build system

component: testsIssues re: TestsIssues re: Tests

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open (in progress).

pytorch/TensorRTnumber 4566

#4566 In pytorch/TensorRT;

· shoumikhin opened on Aug 23, 2026

Ship only the TensorRT delegate in the ExecuTorch runtime wheel

ci: nightlyRun the nightly lane (all tiers incl. llm / kernels / distributed) on every pushRun the nightly lane (all tiers incl. llm / kernels / distributed) on every push

cla signed

component: api [C++]Issues re: C++ APIIssues re: C++ API

component: api [Python]Issues re: Python APIIssues re: Python API

component: build systemIssues re: Build systemIssues re: Build system

component: testsIssues re: TestsIssues re: Tests

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open (in progress).

pytorch/TensorRTnumber 4567

#4567 In pytorch/TensorRT;

· shoumikhin opened on Aug 23, 2026