diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/lit.local.cfg b/mlir/test/Integration/Dialect/SparseTensor/CPU/lit.local.cfg
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/lit.local.cfg
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/lit.local.cfg
@@ -3,3 +3,31 @@
 # No JIT on win32.
 if sys.platform == 'win32':
     config.unsupported = True
+
+# ArmSVE tests must be enabled via build flag.
+if config.mlir_run_arm_sve_tests == 'ON':
+    config.substitutions.append(('%ENABLE_VLA', 'true'))
+    config.substitutions.append(('%VLA_ARCH_ATTR_OPTIONS', '--march=aarch64 --mattr="+sve"'))
+    lli_cmd = 'lli'
+    if config.arm_emulator_lli_executable:
+       lli_cmd = config.arm_emulator_lli_executable
+
+    if config.arm_emulator_utils_lib_dir:
+        config.substitutions.append(('%mlir_native_utils_lib_dir', config.arm_emulator_utils_lib_dir))
+    else:
+        config.substitutions.append(('%mlir_native_utils_lib_dir', config.mlir_integration_test_dir))
+
+    if config.arm_emulator_executable:
+        # Run test in emulator (qemu or armie)
+        emulation_cmd = config.arm_emulator_executable
+        if config.arm_emulator_options:
+            emulation_cmd = emulation_cmd + ' ' + config.arm_emulator_options
+        emulation_cmd = emulation_cmd + ' ' + lli_cmd
+        config.substitutions.append(('%lli', emulation_cmd))
+    else:
+        config.substitutions.append(('%lli', lli_cmd))
+else:
+    config.substitutions.append(('%lli', 'lli'))
+    config.substitutions.append(('%mlir_native_utils_lib_dir', config.mlir_integration_test_dir))
+    config.substitutions.append(('%ENABLE_VLA', 'false'))
+    config.substitutions.append(('%VLA_ARCH_ATTR_OPTIONS', ''))
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_cast.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_cast.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_cast.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_cast.mlir
@@ -11,6 +11,13 @@
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=2 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 #SV = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_filter_conv2d.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_filter_conv2d.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_filter_conv2d.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_filter_conv2d.mlir
@@ -9,6 +9,13 @@
 // RUN: mlir-cpu-runner -e entry -entry-point-result=void \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=2 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 #DCSR = #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ] }>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_flatten.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_flatten.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_flatten.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_flatten.mlir
@@ -13,6 +13,14 @@
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=4 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: TENSOR0="%mlir_integration_test_dir/data/test.tns" \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 !Filename = !llvm.ptr<i8>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_index_dense.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_index_dense.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_index_dense.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_index_dense.mlir
@@ -9,6 +9,13 @@
 // RUN: mlir-cpu-runner -e entry -entry-point-result=void \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=4 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 #SparseVector = #sparse_tensor.encoding<{
   dimLevelType = ["compressed"]
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matvec.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matvec.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matvec.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matvec.mlir
@@ -14,6 +14,14 @@
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=16 enable-vla-vectorization=%ENABLE_VLA enable-simd-index32" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: TENSOR0="%mlir_integration_test_dir/data/wide.mtx" \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 !Filename = !llvm.ptr<i8>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_mttkrp.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_mttkrp.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_mttkrp.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_mttkrp.mlir
@@ -13,6 +13,14 @@
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=4 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: TENSOR0="%mlir_integration_test_dir/data/mttkrp_b.tns" \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 !Filename = !llvm.ptr<i8>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_out_simple.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_out_simple.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_out_simple.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_out_simple.mlir
@@ -13,6 +13,14 @@
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=4 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: TENSOR0="%mlir_integration_test_dir/data/test.mtx" \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 !Filename = !llvm.ptr<i8>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_quantized_matmul.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_quantized_matmul.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_quantized_matmul.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_quantized_matmul.mlir
@@ -9,6 +9,13 @@
 // RUN: mlir-cpu-runner -e entry -entry-point-result=void \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=2 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 #DCSR = #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ] }>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir
@@ -9,6 +9,7 @@
 // RUN: mlir-cpu-runner -e entry -entry-point-result=void \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
 
 #SV = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>
 #DV = #sparse_tensor.encoding<{ dimLevelType = [ "dense"      ] }>
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions_vla.mlir
copy from mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir
copy to mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions_vla.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions_vla.mlir
@@ -1,15 +1,13 @@
-// RUN: mlir-opt %s --sparse-compiler | \
-// RUN: mlir-cpu-runner -e entry -entry-point-result=void \
-// RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
-// RUN: FileCheck %s
 //
-// Do the same run, but now with SIMDization as well. This should not change the outcome.
+// If SVE is available, test VLA vectorization.
 //
-// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=8" | \
-// RUN: mlir-cpu-runner -e entry -entry-point-result=void \
-// RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=8 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
 
+/
+
 #SV = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>
 #DV = #sparse_tensor.encoding<{ dimLevelType = [ "dense"      ] }>
 
@@ -49,30 +47,6 @@
     return %0 : tensor<f32>
   }
 
-  func.func @prod_reduction_i32(%arga: tensor<32xi32, #DV>,
-                           %argx: tensor<i32>) -> tensor<i32> {
-    %0 = linalg.generic #trait_reduction
-      ins(%arga: tensor<32xi32, #DV>)
-      outs(%argx: tensor<i32>) {
-        ^bb(%a: i32, %x: i32):
-          %0 = arith.muli %x, %a : i32
-          linalg.yield %0 : i32
-    } -> tensor<i32>
-    return %0 : tensor<i32>
-  }
-
-  func.func @prod_reduction_f32(%arga: tensor<32xf32, #DV>,
-                           %argx: tensor<f32>) -> tensor<f32> {
-    %0 = linalg.generic #trait_reduction
-      ins(%arga: tensor<32xf32, #DV>)
-      outs(%argx: tensor<f32>) {
-        ^bb(%a: f32, %x: f32):
-          %0 = arith.mulf %x, %a : f32
-          linalg.yield %0 : f32
-    } -> tensor<f32>
-    return %0 : tensor<f32>
-  }
-
   func.func @and_reduction_i32(%arga: tensor<32xi32, #DV>,
                           %argx: tensor<i32>) -> tensor<i32> {
     %0 = linalg.generic #trait_reduction
@@ -164,10 +138,6 @@
        : (tensor<32xi32, #SV>, tensor<i32>) -> tensor<i32>
     %1 = call @sum_reduction_f32(%sparse_input_f32, %rf)
        : (tensor<32xf32, #SV>, tensor<f32>) -> tensor<f32>
-    %2 = call @prod_reduction_i32(%dense_input_i32, %ri)
-       : (tensor<32xi32, #DV>, tensor<i32>) -> tensor<i32>
-    %3 = call @prod_reduction_f32(%dense_input_f32, %rf)
-       : (tensor<32xf32, #DV>, tensor<f32>) -> tensor<f32>
     %4 = call @and_reduction_i32(%dense_input_i32, %ri)
        : (tensor<32xi32, #DV>, tensor<i32>) -> tensor<i32>
     %5 = call @or_reduction_i32(%sparse_input_i32, %ri)
@@ -179,8 +149,6 @@
     //
     // CHECK: 26
     // CHECK: 27.5
-    // CHECK: 3087
-    // CHECK: 168
     // CHECK: 1
     // CHECK: 15
     // CHECK: 10
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_matmul.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_matmul.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_matmul.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_matmul.mlir
@@ -15,6 +15,13 @@
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
 //
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=4 enable-vla-vectorization=%ENABLE_VLA enable-simd-index32" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: TENSOR0="%mlir_integration_test_dir/data/test.mtx" \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 !Filename = !llvm.ptr<i8>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_mm_fusion.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_mm_fusion.mlir
old mode 100755
new mode 100644
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_mm_fusion.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_mm_fusion.mlir
@@ -9,6 +9,13 @@
 // RUN: mlir-cpu-runner -e entry -entry-point-result=void \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=8 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 #SM = #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ] }>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_scale.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_scale.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_scale.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_scale.mlir
@@ -11,6 +11,13 @@
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=4 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 #CSR = #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ] }>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_spmm.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_spmm.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_spmm.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_spmm.mlir
@@ -13,6 +13,14 @@
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=2 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: TENSOR0="%mlir_integration_test_dir/data/wide.mtx" \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 !Filename = !llvm.ptr<i8>
 
diff --git a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum.mlir b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum.mlir
--- a/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum.mlir
+++ b/mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum.mlir
@@ -13,6 +13,14 @@
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
+//
+// If SVE is available, test VLA vectorization.
+//
+// RUN: mlir-opt %s --sparse-compiler="vectorization-strategy=2 vl=2 enable-vla-vectorization=%ENABLE_VLA" | \
+// RUN: mlir-translate -mlir-to-llvmir | \
+// RUN: TENSOR0="%mlir_integration_test_dir/data/test_symmetric.mtx" \
+// RUN: %lli --entry-function=entry %VLA_ARCH_ATTR_OPTIONS --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: FileCheck %s
 
 !Filename = !llvm.ptr<i8>