commit | a884b93c21e02cec786a9a83e773288f3a354413 | [log] [tgz] |
---|---|---|
author | Stanley Winata <68087699+raikonenfnu@users.noreply.github.com> | Fri Aug 16 17:48:22 2024 -0700 |
committer | GitHub <noreply@github.com> | Fri Aug 16 17:48:22 2024 -0700 |
tree | 3525de82966d805daf726e0770f9300f6bce4a97 | |
parent | 0e152d2760abedf29d1f718dbc9754fd47ec77ab [diff] |
Bump LLVM to llvm/llvm-project@ddda37a (#18258) Bump LLVM upstream while reverting https://github.com/llvm/llvm-project/pull/104630 due to issues which will be detailed below. Potential fixes: https://github.com/llvm/llvm-project/pull/104630 for issue no.1 https://github.com/llvm/llvm-project/pull/104648 for issue no.2 https://github.com/llvm/llvm-project/pull/104630 breaks 2 things: 1.vm.call + emitc to c code generation. ```mlir error: null operand found vm.check.eq %ref_dno, %res, "_v_r()=NULL" : !vm.ref<?> %11 = "emitc.call_opaque"(%8, <<NULL VALUE>>) <{callee = "vm_cmp_eq_ref"}> : (!emitc.ptr<!emitc.opaque<"iree_vm_ref_t">>, <<NULL TYPE>>) -> i32 ``` 2. memref_util_test failure due to populateMemrefUtilPatterns not properly legalizing `builtin.unrealized_conversion_cast`, even though we expected `builtin.unrealized_conversion_cast`. ```mlir memref_util_test.mlir:1:79: error: failed to legalize unresolved materialization from ('i32') to 'i16' that remained live after conversion util.func @load_store_i16(%buffer: memref<?xi16>, %idx0: index, %idx1: index, %value: i16) -> i16 { ^ memref_util_test.mlir:1:79: note: see current operation: %0 = "builtin.unrealized_conversion_cast"(%arg3) : (i32) -> i16 memref_util_test.mlir:7:3: note: see existing live user here: "util.buffer.store"(%0, %arg0, %4, %6, %5) : (i16, !util.buffer, index, index, index) -> () memref.store %value, %buffer[%idx0] : memref<?xi16> ^ memref_util_test.mlir:0:0: error: conversion to util failed memref_util_test.mlir:0:0: note: see current operation: "builtin.module"() ({ "util.func"() <{function_type = (memref<?xi16>, index, index, i16) -> i16, sym_name = "load_store_i16", tied_operands = [-1 : index]}> ({ ^bb0(%arg0: memref<?xi16>, %arg1: index, %arg2: index, %arg3: i16): "memref.store"(%arg3, %arg0, %arg1) : (i16, memref<?xi16>, index) -> () %0 = "memref.load"(%arg0, %arg2) : (memref<?xi16>, index) -> i16 "util.return"(%0) : (i16) -> () }) : () -> () }) : () -> () ``` --------- Signed-off-by: Stanley Winata <stanley@nod-labs.com>
IREE (Intermediate Representation Execution Environment, pronounced as “eerie”) is an MLIR-based end-to-end compiler and runtime that lowers Machine Learning (ML) models to a unified IR that scales up to meet the needs of the datacenter and down to satisfy the constraints and special considerations of mobile and edge deployments.
See our website for project details, user guides, and instructions on building from source.
IREE is still in its early phase. We have settled down on the overarching infrastructure and are actively improving various software components as well as project logistics. It is still quite far from ready for everyday use and is made available without any support at the moment. With that said, we welcome any kind of feedback on any communication channels
Package | Release status |
---|---|
GitHub release (stable) | |
GitHub release (nightly) | |
Python iree-compiler | |
Python iree-runtime |
Host platform | Build status |
---|---|
Linux | |
macOS | |
Windows |
For the full list of workflows see https://iree.dev/developers/general/github-actions/.
See our website for more information.
Community meeting recordings: IREE YouTube channel
IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.