sw:vec_iree: Add memory usage log on simple_vec_mul sample

Print memory usage in test_mlexecute

```
14:21:43.3816 [INFO] uart1: [output] KATA> test_mlexecute
14:21:43.4099 [WARNING] cpu2: The debug mode now has no effect - connect
a debugger, and switch to stepping mode.
14:21:43.5901 [INFO] cpu2: simprint: "[[ iree_hal_allocator_t memory
statistics ]]", 0 (0x0)
14:21:43.5908 [INFO] cpu2: simprint: "~  HOST_LOCAL:            0B peak
/            0B allocated /            0B freed /            0B live", 0
(0x0)
14:21:43.5909 [INFO] cpu2: simprint: "DEVICE_LOCAL:         4096B peak /
4096B allocated /         4096B freed /            0B live", 0 (0x0)
14:21:43.5926 [INFO] cpu2: simprint: "INFO |simple_vec_mul finished
successfully", 0 (0x0)
14:21:43.5936 [INFO] cpu2: simprint: "main returned: ", 0 (0x0)
```

Due to b/195166567, only the first execution has the memory log printed
(IREE internal print mechanism uses fprintf).

Change-Id: I0d5f0abec085fcdf6604276c4c3692d6fa97084d

samples/simple_vec_mul/simple_vec_mul.c

diff --git a/samples/simple_vec_mul/simple_vec_mul.c b/samples/simple_vec_mul/simple_vec_mul.c
index 4bd6dc3..c32d3cf 100644
--- a/samples/simple_vec_mul/simple_vec_mul.c
+++ b/samples/simple_vec_mul/simple_vec_mul.c
@@ -146,8 +146,10 @@
   iree_vm_list_release(outputs);
   iree_aligned_free(arg0_buffer);
   iree_aligned_free(arg1_buffer);
-  iree_hal_device_release(device);
   iree_vm_context_release(context);
+  IREE_IGNORE_ERROR(iree_hal_allocator_statistics_fprint(
+      stdout, iree_hal_device_allocator(device)));
+  iree_hal_device_release(device);
   iree_vm_instance_release(instance);
   return result;
 }