[mlir][test][Integration] Refactor Arm emulator configuration

[c-rhodes]

Summary

The logic enabling the Arm SVE (and now SME) integration tests for
various dialects, that may run under emulation, is now duplicated in
several places.

This patch moves the configuration to the top-level MLIR integration
tests Lit config and renames the '%lli' substitution in contexts where
it will run exclusively (ArmSVE, ArmSME) on AArch64 (and possibly under
emulation) to '%lli_aarch64_cmd', and '%lli_host_or_aarch64_cmd' for
contexts where it will run on the host OR AArch64 (also possibly under
emulation). The latter is for integration tests that have
target-specific and target-agnostic codepaths such as SparseTensor,
which supports scalable vectors.

The two substitutions have the same effect but the names are different to
convey this information. The '%lli_aarch64_cmd' substitution could be
used in the SparseTensor tests but that would be a misnomer if the host
were x86 and the MLIR_RUN_SVE_TESTS=OFF.

The reason for renaming the '%lli' substitution is to not prevent running other
target-specific integration tests at the same time, since the same substitution
'%lli' is used for lli in other integration tests:

• mlir/test/Integration/Dialect/Vector/CPU/X86Vector - (AVX emulation via Intel SDE)
• mlir/test/Integration/Dialect/Vector/CPU/AMX - (AMX emulation via Intel SDE)
• mlir/test/Integration/Dialect/LLVMIR/CPU/test-vp-intrinsic.mlir - (RISCV emulation via QEMU if supported, native otherwise)

and substituting '%lli' at the top-level with Arm specific logic would override
this.

[awarzynski]

Many thanks for preparing this, @c-rhodes ! I am very much in favor of replacing generic LIT variables like %lli (that change meaning depending on context) with some more explicit variables. Also, it's always nice to see code being removed :)

These changes make a lot of sense to, though it seems there's room for a couple more improvements. Could you take a look? Ta!

mlir/test/Integration/Dialect/SparseTensor/CPU/lit.local.cfg

15–16

It feels that these two could be moved to "mlir/test/Integration/lit.local.cfg" too? This way, this config would only contain things "specific" to SparseTensor tests (these two LIT variables are more generic than that).

Otherwise, it would be good to clarify the relationship between this and what's in "mlir/test/Integration/lit.local.cfg" .

mlir/test/Integration/lit.local.cfg

14–15

[nit] Move "unconditional" code to the top of this block (because this applies to all configs, irrespective of the value of e.g. arm_emulator_lli_executable)

17

This would make a bit more sense to me:

if not config.arm_emulator_executable:
    config.substitutions.append(('%lli_aarch64_cmd', lli_cmd))
    config.substitutions.append(('%lli_host_or_aarch64_cmd', lli_cmd))
else
  (...)

This way, it's immediately obvious that the purpose of these two blocks is to set-up substitutions for '%lli_aarch64_cmd and %lli_host_or_aarch64_cmd. Otherwise that's obfuscated a bit.

31

1. These two substitutions are equivalent

I think that this applies to the whole block, because you basically have:

if config.mlir_run_arm_sve_tests or config.mlir_run_arm_sme_tests:
        config.substitutions.append(('%lli_aarch64_cmd', emulation_cmd))
        config.substitutions.append(('%lli_host_or_aarch64_cmd', emulation_cmd))
    else:
        config.substitutions.append(('%lli_aarch64_cmd', lli_cmd))
        config.substitutions.append(('%lli_host_or_aarch64_cmd', lli_cmd))

So I would move this particular comment and also clarify, e.g.:

This block configures substitutions for:
*  `%lli_aarch64_cmd`, and 
* `%lli_host_or_aarch64_cmd'`. 

The former is only relevant/required when SVE integrations tests are enabled. The latter changes meaning depending on whether those tests are enabled. When the SVE tests _are_ enabled then both are equivalent.

[aartbik]

This is absolutely LGTM for the sparse part, thanks for following up
(I leave the arms specific review to the experts, but the sparse tests look very clean now)

[c-rhodes]

Move '%lli_host_or_aarch64_cmd' -> 'lli' substitution from SparseTensor Lit config to top-level and move logic to function.

[c-rhodes]

In D148929#4292386, @awarzynski wrote:

Many thanks for preparing this, @c-rhodes ! I am very much in favor of replacing generic LIT variables like %lli (that change meaning depending on context) with some more explicit variables. Also, it's always nice to see code being removed :)

These changes make a lot of sense to, though it seems there's room for a couple more improvements. Could you take a look? Ta!

Cheers Andrzej, updated.

mlir/test/Integration/Dialect/SparseTensor/CPU/lit.local.cfg

15–16

It feels that these two could be moved to "mlir/test/Integration/lit.local.cfg" too? This way, this config would only contain things "specific" to SparseTensor tests (these two LIT variables are more generic than that).

Otherwise, it would be good to clarify the relationship between this and what's in "mlir/test/Integration/lit.local.cfg" .

I've moved '%lli_host_or_aarch64_cmd' -> 'lli' to top-level. %mlir_native_utils_lib_dir is also used by RISCV in mlir/test/Integration/Dialect/LLVMIR/CPU/test-vp-intrinsic.mlir so moving this substitution to top-level and predicating on !config.mlir_run_arm_sve_tests would break that test unless additional logic was added to handle that, so I think it's simpler to keep it here.

mlir/test/Integration/lit.local.cfg

17

This would make a bit more sense to me:

if not config.arm_emulator_executable:
    config.substitutions.append(('%lli_aarch64_cmd', lli_cmd))
    config.substitutions.append(('%lli_host_or_aarch64_cmd', lli_cmd))
else
  (...)

This way, it's immediately obvious that the purpose of these two blocks is to set-up substitutions for '%lli_aarch64_cmd and %lli_host_or_aarch64_cmd. Otherwise that's obfuscated a bit.

I've refactored this into a function, I think it's cleaner now and addresses your comments, let me know if not.

[awarzynski]

LGTM, this is a massive improvement, thanks!

mlir/test/Integration/lit.local.cfg

38–42

[nit] I think that splitting this into bullet points will improve readability (that's the only change that I am suggesting)

[c-rhodes]

In D148929#4298109, @awarzynski wrote:

LGTM, this is a massive improvement, thanks!

Cheers Andrzej. I'll address your nit before committing.

mlir/test/Integration/Dialect/SparseTensor/CPU/concatenate_dim_0.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_lib_dir/libmlir_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #MAT_C_C = #sparse_tensor.encoding<{dimLevelType = ["compressed", "compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/concatenate_dim_0_permute.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_lib_dir/libmlir_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #MAT_C_C = #sparse_tensor.encoding<{dimLevelType = ["compressed", "compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/concatenate_dim_1_permute.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_lib_dir/libmlir_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #MAT_C_C = #sparse_tensor.encoding<{dimLevelType = ["compressed", "compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/dense_output.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/test.mtx" \
-// REDEFINE: %lli \
+// REDEFINE: %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_runner_utils | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 !Filename = !llvm.ptr<i8>

mlir/test/Integration/Dialect/SparseTensor/CPU/dense_output_f16.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/lit.local.cfg

 import sys
 from lit.llvm import llvm_config
 
 # FIXME: %mlir_native_utils_lib_dir is set incorrectly on Windows
 if sys.platform == 'win32':
     config.unsupported = True
 
 # ArmSVE tests must be enabled via build flag.
 if config.mlir_run_arm_sve_tests:
     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_lib_dir))
-
-    if config.arm_emulator_executable:
-        if not config.arm_emulator_lli_executable:
-            # Top-level lit config adds llvm_tools_dir to PATH but this is lost
-            # when running under an emulator. If the user didn't specify an lli
-            # executable, use absolute path %llvm_tools_dir/lli.
-            # TODO(c-rhodes): This logic is duplicated across several Lit files
-            # and needs refactoring.
-            lli_cmd = llvm_config.use_llvm_tool('lli', search_env='LLI', required=True,
-                                                search_paths=[config.llvm_tools_dir],
-                                                use_installed=False)
-        # 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(('%ENABLE_VLA', 'false'))
     config.substitutions.append(('%VLA_ARCH_ATTR_OPTIONS', ''))
-    config.substitutions.append(('%lli', 'lli'))
+    config.substitutions.append(('%lli_host_or_aarch64_cmd', 'lli'))
     config.substitutions.append(('%mlir_native_utils_lib_dir', config.mlir_lib_dir))

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_abs.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_binary.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_cast.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SV = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_codegen_dim.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR  = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_codegen_foreach.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #Row = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_collapse_shape.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_complex32.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_complex64.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_complex_ops.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_constant_to_sparse_tensor.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #Tensor1  = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conv_1d_nwc_wcf.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #CCC = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conv_2d.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR = #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conv_2d_nhwc_hwcf.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #CCCC = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conv_3d.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #CCC = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conv_3d_ndhwc_dhwcf.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #CCCCC = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conversion.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #Tensor1  = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conversion_dyn.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 reassociate-fp-reductions=true enable-index-optimizations=true enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_runner_utils | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR  = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conversion_element.mlir

 // DEFINE: %{option} = "enable-runtime-library=false s2s-strategy=2"
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false s2s-strategy=2 vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #Tensor1 = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conversion_ptr.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 reassociate-fp-reductions=true enable-index-optimizations=true enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR  = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conversion_sparse2dense.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #Tensor1  = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conversion_sparse2sparse.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false s2s-strategy=2 vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #Tensor1 = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_coo_test.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SortedCOO = #sparse_tensor.encoding<{
func.func @entry() {
     bufferization.dealloc_tensor %CSR_A : tensor<8x8xf32, #CSR>
     bufferization.dealloc_tensor %COO_A : tensor<8x8xf32, #SortedCOO>
     bufferization.dealloc_tensor %COO_B : tensor<8x8xf32, #SortedCOO>
 
 
     return
   }
 }
 No newline at end of file

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_dot.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_expand.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_runner_utils | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #CSC = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_expand_shape.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_file_io.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/wide.mtx" TENSOR1="" \
-// REDEFINE:   %lli \
+// REDEFINE:   %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 !Filename = !llvm.ptr<i8>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_filter_conv2d.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR = #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_flatten.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/test.tns" \
-// REDEFINE: %lli \
+// REDEFINE: %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_runner_utils | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 !Filename = !llvm.ptr<i8>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_index.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_index_dense.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_insert_1d.mlir

 // DEFINE: %{option} = enable-runtime-library=false
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 // Insertion example using pure codegen (no sparse runtime support lib).

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_insert_2d.mlir

 // DEFINE: %{option} = enable-runtime-library=false
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #Dense = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_insert_3d.mlir

 // DEFINE: %{option} = enable-runtime-library=false
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #TensorCSR = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matmul.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_runner_utils | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 // TODO: Investigate the output generated for SVE, see https://github.com/llvm/llvm-project/issues/60626

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matrix_ops.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_runner_utils | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR = #sparse_tensor.encoding<{dimLevelType = ["compressed", "compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matvec.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/wide.mtx" \
-// REDEFINE: %lli \
+// REDEFINE: %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 !Filename = !llvm.ptr<i8>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_mttkrp.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4  enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/mttkrp_b.tns" \
-// REDEFINE: %lli \
+// REDEFINE: %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext -dlopen=%mlir_runner_utils| \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 !Filename = !llvm.ptr<i8>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_out_mult_elt.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_out_reduction.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseMatrix = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_out_simple.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4  enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/test.mtx" \
-// REDEFINE: %lli \
+// REDEFINE: %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 !Filename = !llvm.ptr<i8>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_pack.mlir

 // DEFINE: %{option} = enable-runtime-library=false
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 // TODO: Pack only support CodeGen Path

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_quantized_matmul.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR = #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_re_im.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reduce_custom.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false  enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 reassociate-fp-reductions=true enable-index-optimizations=true enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 // Reduction in this file _are_ supported by the AArch64 SVE backend

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 
 // Reduction in this file _are_ supported by the AArch64 SVE backend
 

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_rewrite_push_back.mlir

 // DEFINE: %{option} = enable-runtime-library=false
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 module {

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_rewrite_sort.mlir

 // DEFINE: %{option} = enable-runtime-library=false
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils,%mlir_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext --dlopen=%mlir_runner_utils | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 module {

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_rewrite_sort_coo.mlir

 // DEFINE: %{option} = enable-runtime-library=false
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 module {

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_matmul.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4  enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/test.mtx" \
-// REDEFINE: %lli \
+// REDEFINE: %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 !Filename = !llvm.ptr<i8>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_mm_fusion.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SM = #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_scale.mlir

 // DEFINE: %{option} = enable-runtime-library=true
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with direct IR generation.
 // REDEFINE: %{option} = enable-runtime-library=false
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #CSR = #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_scf_nested.mlir

 // DEFINE: %{option} = enable-runtime-library=true
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with direct IR generation.
 // REDEFINE: %{option} = enable-runtime-library=false
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #map = affine_map<(d0, d1, d2) -> (d0, d1, d2)>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_select.mlir

 // DEFINE: %{option} = enable-runtime-library=true
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with direct IR generation.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sign.mlir

 // DEFINE: %{option} = enable-runtime-library=true
 // DEFINE: %{compile} = mlir-opt %s --sparse-compiler=%{option}
 // DEFINE: %{run} = mlir-cpu-runner \
 // DEFINE:  -e entry -entry-point-result=void  \
 // DEFINE:  -shared-libs=%mlir_c_runner_utils | \
 // DEFINE: FileCheck %s
 //
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with direct IR generation.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_storage.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 //

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4  enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/test_symmetric.mtx" \
-// REDEFINE: %lli \
+// REDEFINE: %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 // TODO: The test currently only operates on the triangular part of the

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum_c32.mlir

 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
 // REDEFINE: %{run} = TENSOR0="%mlir_src_dir/test/Integration/data/test_symmetric_complex.mtx" \
-// REDEFINE:   %lli \
+// REDEFINE:   %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 // TODO: The test currently only operates on the triangular part of the

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum_f16.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 !Filename = !llvm.ptr<i8>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_tanh.mlir

 // RUN: %{compile} | %{run}
 
 // Current fails for SVE, see https://github.com/llvm/llvm-project/issues/60626
 // UNSUPPORTED: target=aarch64{{.*}}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 reassociate-fp-reductions=true enable-index-optimizations=true enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_tensor_mul.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #ST = #sparse_tensor.encoding<{dimLevelType = ["compressed", "compressed", "compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_tensor_ops.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #ST1 = #sparse_tensor.encoding<{dimLevelType = ["compressed", "compressed", "compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_transpose.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #DCSR = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_transpose_coo.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 //
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SortedCOO = #sparse_tensor.encoding<{

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_unary.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false vl=4 enable-buffer-initialization=true reassociate-fp-reductions=true enable-index-optimizations=true enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_vector_ops.mlir

 //
 // Do the same run, but now with direct IR generation and vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=2 reassociate-fp-reductions=true enable-index-optimizations=true"
 // RUN: %{compile} | %{run}
 
 // Do the same run, but now with direct IR generation and, if available, VLA
 // vectorization.
 // REDEFINE: %{option} = "enable-runtime-library=false enable-buffer-initialization=true vl=4 reassociate-fp-reductions=true enable-index-optimizations=true enable-arm-sve=%ENABLE_VLA"
-// REDEFINE: %{run} = %lli \
+// REDEFINE: %{run} = %lli_host_or_aarch64_cmd \
 // REDEFINE:   --entry-function=entry_lli \
 // REDEFINE:   --extra-module=%S/Inputs/main_for_lli.ll \
 // REDEFINE:   %VLA_ARCH_ATTR_OPTIONS \
 // REDEFINE:   --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // REDEFINE: FileCheck %s
 // RUN: %{compile} | mlir-translate -mlir-to-llvmir | %{run}
 
 #SparseVector = #sparse_tensor.encoding<{dimLevelType = ["compressed"]}>

mlir/test/Integration/Dialect/Vector/CPU/ArmSME/lit.local.cfg

 import sys
-from lit.llvm import llvm_config
 
 # ArmSME tests must be enabled via build flag.
 if not config.mlir_run_arm_sme_tests:
     config.unsupported = True
 
 # No JIT on win32.
 if sys.platform == 'win32':
     config.unsupported = True
-
-lli_cmd = 'lli'
-if config.arm_emulator_lli_executable:
-    lli_cmd = config.arm_emulator_lli_executable
-
-config.substitutions.append(('%mlir_native_utils_lib_dir',
-    config.arm_emulator_utils_lib_dir or config.mlir_lib_dir))
-
-if config.arm_emulator_executable:
-    if not config.arm_emulator_lli_executable:
-        # Top-level lit config adds llvm_tools_dir to PATH but this is lost
-        # when running under an emulator. If the user didn't specify an lli
-        # executable, use absolute path %llvm_tools_dir/lli.
-        # TODO(c-rhodes): This logic is duplicated across several Lit files and
-        # needs refactoring.
-        lli_cmd = llvm_config.use_llvm_tool('lli', search_env='LLI', required=True,
-                                            search_paths=[config.llvm_tools_dir],
-                                            use_installed=False)
-    # Run test in emulator (QEMU)
-    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))

mlir/test/Integration/Dialect/Vector/CPU/ArmSME/test-ssve.mlir

 // RUN: mlir-opt %s -test-lower-to-llvm | \
 // RUN: mlir-translate -mlir-to-llvmir | \
-// RUN: %lli --march=aarch64 --mattr="+sve,+sme" \
+// RUN: %lli_aarch64_cmd --march=aarch64 --mattr="+sve,+sme" \
 // RUN:      -force-streaming-compatible-sve \
 // RUN:      --entry-function=entry \
 // RUN:      --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
 
 // NOTE: To run this test, your CPU must support SME.
 
 // VLA memcopy in streaming mode.

mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/lit.local.cfg

 import sys
-from lit.llvm import llvm_config
 
 # ArmSVE tests must be enabled via build flag.
 if not config.mlir_run_arm_sve_tests:
     config.unsupported = True
 
 # No JIT on win32.
 if sys.platform == 'win32':
     config.unsupported = True
-
-lli_cmd = 'lli'
-if config.arm_emulator_lli_executable:
-    lli_cmd = config.arm_emulator_lli_executable
-
-config.substitutions.append(('%mlir_native_utils_lib_dir',
-    config.arm_emulator_utils_lib_dir or config.mlir_lib_dir))
-
-if config.arm_emulator_executable:
-    if not config.arm_emulator_lli_executable:
-        # Top-level lit config adds llvm_tools_dir to PATH but this is lost
-        # when running under an emulator. If the user didn't specify an lli
-        # executable, use absolute path %llvm_tools_dir/lli.
-        # TODO(c-rhodes): This logic is duplicated across several Lit files and
-        # needs refactoring.
-        lli_cmd = llvm_config.use_llvm_tool('lli', search_env='LLI', required=True,
-                                            search_paths=[config.llvm_tools_dir],
-                                            use_installed=False)
-    # 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))

mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/test-sve.mlir

 // RUN: mlir-opt %s -lower-affine -convert-scf-to-cf -convert-vector-to-llvm="enable-arm-sve" -finalize-memref-to-llvm -convert-func-to-llvm -convert-arith-to-llvm -canonicalize | \
 // RUN: mlir-translate -mlir-to-llvmir | \
-// RUN: %lli --entry-function=entry --march=aarch64 --mattr="+sve" --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
+// RUN: %lli_aarch64_cmd --entry-function=entry --march=aarch64 --mattr="+sve" --dlopen=%mlir_native_utils_lib_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
 
 // Note: To run this test, your CPU must support SVE
 
 // VLA memcopy
 func.func @kernel_copy(%src : memref<?xi64>, %dst : memref<?xi64>, %size : index) {
   %c0 = arith.constant 0 : index
   %c2 = arith.constant 2 : index

mlir/test/Integration/lit.local.cfg

+from lit.llvm import llvm_config
+
 if not config.mlir_include_integration_tests:
     config.unsupported = True
+
+# NOTE: If the SVE tests are disabled and the SME tests are enabled to run
+# under emulation, the SVE specific RUN lines in the SparseTensor tests will
+# run under emulation.
+if config.mlir_run_arm_sve_tests or config.mlir_run_arm_sme_tests:
+    lli_cmd = 'lli'
+    if config.arm_emulator_lli_executable:
+        lli_cmd = config.arm_emulator_lli_executable
+
+    config.substitutions.append(('%mlir_native_utils_lib_dir',
+        config.arm_emulator_utils_lib_dir or config.mlir_lib_dir))
+
+    if config.arm_emulator_executable:
+        if not config.arm_emulator_lli_executable:
+            # Top-level lit config adds llvm_tools_dir to PATH but this is lost
+            # when running under an emulator. If the user didn't specify an lli
+            # executable, use absolute path %llvm_tools_dir/lli.
+            lli_cmd = llvm_config.use_llvm_tool(
+                'lli', search_env='LLI', required=True,
+                search_paths=[config.llvm_tools_dir], use_installed=False
+            )
+        # 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
+        # These two substitutions are equivalent but '%lli_aarch64_cmd' should
+        # be used for integration tests that run exclusively on AArch64
+        # (ArmSVE, ArmSME), and '%lli_host_or_aarch64_cmd' should be used for
+        # tests that may run on the host OR AArch64, such as the SparseTensor
+        # tests which have target-specific (SVE) and target-agnostic codepaths.
+        config.substitutions.append(('%lli_aarch64_cmd', emulation_cmd))
+        config.substitutions.append(('%lli_host_or_aarch64_cmd', emulation_cmd))
+    else:
+        config.substitutions.append(('%lli_aarch64_cmd', lli_cmd))
+        config.substitutions.append(('%lli_host_or_aarch64_cmd', lli_cmd))