[TOPI]fix scatterND large shape problem (apache#12200)

* fix scatterND large shape problem

* fix thread pool alloca

* add scatternd unit test

* update with comment

* Empty

Co-authored-by: wrongtest <wrongtest0@gmail.com>

src/target/llvm/codegen_cpu.cc

@@ -642,7 +642,8 @@ CodeGenLLVM::TypedPointer CodeGenCPU::PackClosureData(const Array<Var>& vfields,
   }
   llvm::StructType* ctype = struct_name.size() ? llvm::StructType::create(fields, struct_name)
                                                : llvm::StructType::create(fields);
-  llvm::Value* cvalue = builder_->CreateAlloca(ctype, ConstInt32(1));
+  llvm::AllocaInst* cvalue =
+      WithFunctionEntry([&]() { return builder_->CreateAlloca(ctype, ConstInt32(1)); });
   llvm::Value* zero = ConstInt32(0);
   for (size_t i = 0; i < vfields.size(); ++i) {
     builder_->CreateStore(var_map_.at(vfields[i].get()),

tests/python/relay/test_op_level3.py

@@ -1909,6 +1909,26 @@ def test_cumprod(target, dev, executor_kind):
 
 @tvm.testing.parametrize_targets
 def test_scatter_nd(target, dev, executor_kind):
+    def test_scatter_nd_large_shape():
+        def before():
+            data = relay.const(np.zeros((1, 900, 300), dtype="float32"), dtype="float32")
+            indices = relay.const(np.ones((3, 1, 900, 300), dtype="int64"), dtype="int64")
+            update = relay.const(np.ones((1, 900, 300), dtype="float32"), dtype="float32")
+            b = relay.op.scatter_nd(data, indices, update)
+            return relay.Function(relay.analysis.free_vars(b), b)
+
+        passes = tvm.transform.Sequential(
+            [
+                relay.transform.InferType(),
+                relay.transform.FoldConstant(),
+            ]
+        )
+        before_mod = tvm.IRModule.from_expr(before())
+        with tvm.transform.PassContext(opt_level=3):
+            after_mod = passes(before_mod)
+
+    test_scatter_nd_large_shape()
+
     def verify_scatter_nd(
         data_np, indices_np, updates_np, ref_res, mode="add", rtol=1e-5, atol=1e-5
     ):