Integrate LLVM at llvm/llvm-project@7752fec6 #18177
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: hanhanW <hanhan0912@gmail.com>
Abbreviated Benchmark Summary@ commit a7057b33695a46ea551c12173d37e9f4bf451106 (vs. base 5a48912c52f65ead960bec9bfde5a836d1b02ab2) Data-Tiling Comparison TableClick to show
No improved or regressed benchmarks 🏖️ No improved or regressed compilation metrics 🏖️ For more information: |
hanhanW
changed the title
Signed-off-by: hanhanW <hanhan0912@gmail.com>
ScottTodd
approved these changes
Aug 9, 2024
It looks a LLVM failure.. looking |
Signed-off-by: hanhanW <hanhan0912@gmail.com>
|
Confirmed that llvm/llvm-project#102198 is the issue locally. I'll try to create a repro and signal it in upstream. |
|
To repro it in IREE: func.func private @generate_2D_source(%height : index, %width : index) -> tensor<?x?xi32> {
%init_source = tensor.empty(%height, %width) : tensor<?x?xi32>
%source = linalg.generic {
indexing_maps = [affine_map<(d0, d1) -> (d0, d1)>],
iterator_types = ["parallel", "parallel"]}
outs(%init_source : tensor<?x?xi32>) {
^bb0(%b0 : i32):
%outer = linalg.index 0 : index
%inner = linalg.index 1 : index
%strided = arith.muli %outer, %width : index
%linearized = arith.addi %inner, %strided : index
%linearized_i32 = arith.index_cast %linearized : index to i32
linalg.yield %linearized_i32 : i32
} -> tensor<?x?xi32>
// This blocks the fusion for inputs and testing ops.
%0 = util.optimization_barrier %source : tensor<?x?xi32>
%1 = flow.tensor.tie_shape %0 : tensor<?x?xi32>{%height, %width}
return %1 : tensor<?x?xi32>
}
func.func @static_pack_pad_transpose_outer_dims_large() {
%height = arith.constant 100 : index
%width = arith.constant 250 : index
%0 = call @generate_2D_source(%height, %width) : (index, index) -> tensor<?x?xi32>
%source = tensor.cast %0 : tensor<?x?xi32> to tensor<100x250xi32>
%padding_value = arith.constant 42 : i32
%init_pack = tensor.empty() : tensor<16x4x32x16xi32>
%pack = tensor.pack %source padding_value(%padding_value : i32)
outer_dims_perm = [1, 0] inner_dims_pos = [0, 1] inner_tiles = [32, 16] into %init_pack
: tensor<100x250xi32> -> tensor<16x4x32x16xi32>
%pad = tensor.pad %source low[0, 0] high[28, 6] {
^bb0(%b0 : index, %b1 : index):
tensor.yield %padding_value : i32
} : tensor<100x250xi32> to tensor<128x256xi32>
%reshape = tensor.expand_shape %pad [[0, 1], [2, 3]] output_shape [4, 32, 16, 16] : tensor<128x256xi32> into tensor<4x32x16x16xi32>
%init_transpose = tensor.empty() : tensor<16x4x32x16xi32>
%transpose = linalg.transpose
ins(%reshape : tensor<4x32x16x16xi32>)
outs(%init_transpose : tensor<16x4x32x16xi32>)
permutation = [2, 0, 1, 3]
check.expect_eq(%pack, %transpose) : tensor<16x4x32x16xi32>
return
} |
|
Here is the LLVM module before translate it to ISA: https://gist.github.com/hanhanW/b806883680dd028ed82943b84483c2b8 And it failed in which calls iree/compiler/plugins/target/ROCM/ROCMTarget.cpp Lines 191 to 203 in 5a48912 @MaheshRavishankar @kuhar how do we reproduce the failure using LLVM tools? I think we can ask for a revert if we can provide such repro to the author. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Revert commits: