Qualcomm AI Engine Direct - heap profiling at runtime on target (pytorch#20272)

### Summary
    Heap profiling at runtime with HTP backend on Android platforms. DSP
heap profiling is available for QnnContext_createFromBinary use-cases.
    It captures total DSP heap usage at two checkpoints:
    - Before the first context is created (before_context_created)
    - After the last context is freed (after_context_freed)

The difference between the two values represents heap consumed during
    context execution. The value after freeing is typically equal to or
    greater than before creation.

    ### Test plan
    python backends/qualcomm/tests/test_qnn_delegate.py
TestQNNQuantizedUtils.test_qnn_backend_runtime_option_heap_profile -b
    build-android -H ${HOST} -s ${SN} -m ${SOC_MODEL}

    ### Note
    This test is expected to run on target device.

Author: jethroqti

backends/qualcomm/tests/test_qnn_delegate.py

@@ -6181,6 +6181,8 @@ def test_qnn_backend_profile_op(self):
         TestQNN.profile_level = 0
 
     def test_qnn_backend_runtime_option_heap_profile(self):
+        if self.enable_x86_64:
+            self.skipTest("heap profiling is not supported on host machine")
         module = SimpleModel()  # noqa: F405
         sample_input = (torch.ones(1, 32, 28, 28), torch.ones(1, 32, 28, 28))
 
@@ -7122,6 +7124,8 @@ def test_qnn_backend_profile_op(self):
         TestQNN.profile_level = 0
 
     def test_qnn_backend_runtime_option_heap_profile(self):
+        if self.enable_x86_64:
+            self.skipTest("heap profiling is not supported on host machine")
         module = SimpleModel()  # noqa: F405
         sample_input = (torch.ones(1, 32, 28, 28), torch.ones(1, 32, 28, 28))
         module1 = self.get_qdq_module(module, sample_input)