Integrate LLVM at llvm/llvm-project@8b4afc5aef14

Updates LLVM usage to match
[8b4afc5aef14](https://github.com/llvm/llvm-project/commit/8b4afc5aef14)

PiperOrigin-RevId: 396535042

SUBMODULE_VERSIONS.txt

diff --git a/SUBMODULE_VERSIONS.txt b/SUBMODULE_VERSIONS.txt
index a5f6132..2c3a066 100644
--- a/SUBMODULE_VERSIONS.txt
+++ b/SUBMODULE_VERSIONS.txt
@@ -4,7 +4,7 @@
 aa533abfd4232b01f9e57041d70114d5a77e6de0 third_party/googletest
 88b845dee001723c4a0db1fe5477de735b6d3bb0 third_party/liburing
 acd6f6f014c25e46363e718381e0b35205df2d83 third_party/libyaml
-3ca6eee2a975edcfa49d8adff6b90df6f8e1ba85 third_party/llvm-project
+8b4afc5aef148aff26047ca7bad4cdcf58c35e25 third_party/llvm-project
 75a20d0492313e5a18eb77d62832ddbe28f0db97 third_party/mlir-hlo
 3f701faace7addc75d16dea8a6cd769fa5b3f260 third_party/musl
 4c7697dbe973ed01ae6fbec37d186ebd05982e1f third_party/pybind11

iree/compiler/Dialect/Flow/Transforms/test/dispatch_linalg_on_tensors_elementwise.mlir

diff --git a/iree/compiler/Dialect/Flow/Transforms/test/dispatch_linalg_on_tensors_elementwise.mlir b/iree/compiler/Dialect/Flow/Transforms/test/dispatch_linalg_on_tensors_elementwise.mlir
index 03c5e5e..466c7e5 100644
--- a/iree/compiler/Dialect/Flow/Transforms/test/dispatch_linalg_on_tensors_elementwise.mlir
+++ b/iree/compiler/Dialect/Flow/Transforms/test/dispatch_linalg_on_tensors_elementwise.mlir
@@ -115,7 +115,7 @@
   return %2 : tensor<1280xf32>
 }
 
-//  CHECK-DAG: #[[SIZE_MAP0:.+]] = affine_map<(d0, d1) -> (d1, -d0 + 1280)>
+//  CHECK-DAG: #[[SIZE_MAP0:.+]] = affine_map<(d0, d1) -> (d0, -d1 + 1280)>
 //  CHECK-DAG: #[[SIZE_MAP1:.+]] = affine_map<(d0, d1) -> (-d0 + 1280, d1)>
 
 //      CHECK: func @tile_parallel_reduction
@@ -127,7 +127,7 @@
 // CHECK-NEXT:     (%[[ARG1:.+]]: !flow.dispatch.tensor<readonly:7x7x1280xf32>, %[[ARG2:.+]]: !flow.dispatch.tensor<writeonly:1280xf32>) {
 //      CHECK:   %[[WG_SIZE0:.+]] = flow.dispatch.workgroup.size[0] : index
 //      CHECK:   scf.for %[[IV:.+]] = %{{.+}} to %{{.+}} step %{{.+}}
-//      CHECK:     %[[SIZE0:.+]] = affine.min #[[SIZE_MAP0]](%[[IV]], %[[WG_SIZE0]])
+//      CHECK:     %[[SIZE0:.+]] = affine.min #[[SIZE_MAP0]](%[[WG_SIZE0]], %[[IV]])
 //      CHECK:     %[[IN:.+]] = flow.dispatch.tensor.load %[[ARG1]],
 // CHECK-SAME:       sizes = [7, 7, %[[SIZE0]]]
 //      CHECK:     %[[SIZE1:.+]] = affine.min #[[SIZE_MAP1]](%[[IV]], %[[WG_SIZE0]])

third_party/llvm-project

diff --git a/third_party/llvm-project b/third_party/llvm-project
index 3ca6eee..8b4afc5 160000
--- a/third_party/llvm-project
+++ b/third_party/llvm-project
@@ -1 +1 @@
-Subproject commit 3ca6eee2a975edcfa49d8adff6b90df6f8e1ba85
+Subproject commit 8b4afc5aef148aff26047ca7bad4cdcf58c35e25