diff --git a/taichi/codegen/spirv/spirv_codegen.cpp b/taichi/codegen/spirv/spirv_codegen.cpp
index 3a15f8d560e6c..4e28cb981c901 100644
--- a/taichi/codegen/spirv/spirv_codegen.cpp
+++ b/taichi/codegen/spirv/spirv_codegen.cpp
@@ -208,6 +208,7 @@ class TaskCodegen : public IRVisitor {
 
       spirv::SType arr_type = ir_->get_array_type(elem_type, elem_num);
       spirv::Value ptr_val = ir_->alloca_workgroup_array(arr_type);
+      shared_array_binds_.push_back(ptr_val);
       ir_->register_value(alloca->raw_name(), ptr_val);
     } else {
       // Alloca for a single variable
@@ -1567,6 +1568,7 @@ class TaskCodegen : public IRVisitor {
     ir_->set_work_group_size(group_size);
     std::vector<spirv::Value> buffers;
     if (device_->get_cap(DeviceCapability::spirv_version) > 0x10300) {
+      buffers = shared_array_binds_;
       for (const auto &bb : task_attribs_.buffer_binds) {
         for (auto &it : buffer_value_map_) {
           if (it.first.first == bb.buffer) {
@@ -2220,6 +2222,7 @@ class TaskCodegen : public IRVisitor {
                      BufferInfoTypeTupleHasher>
       buffer_binding_map_;
   std::vector<TextureBind> texture_binds_;
+  std::vector<spirv::Value> shared_array_binds_;
   spirv::Value kernel_function_;
   spirv::Label kernel_return_label_;
   bool gen_label_{false};
diff --git a/tests/python/test_shared_array.py b/tests/python/test_shared_array.py
index 53b129d6e7b45..64a9a60f7383b 100644
--- a/tests/python/test_shared_array.py
+++ b/tests/python/test_shared_array.py
@@ -4,7 +4,7 @@
 from tests import test_utils
 
 
-@test_utils.test(arch=[ti.cuda, ti.vulkan], vk_api_version="1.0")
+@test_utils.test(arch=[ti.cuda, ti.vulkan])
 def test_shared_array_nested_loop():
     block_dim = 128
     nBlocks = 64