Tensor: dispose Tensors that have been garbage collected (thread safe) + test

src/TensorFlowNET.Core/Status/Status.cs

@@ -26,7 +26,7 @@ namespace Tensorflow
     /// </summary>
     public class Status : IDisposable
     {
-        protected readonly IntPtr _handle;
+        protected IntPtr _handle;
 
         /// <summary>
         /// Error message
@@ -71,7 +71,20 @@ namespace Tensorflow
 
         public void Dispose()
         {
-            c_api.TF_DeleteStatus(_handle);
+            IntPtr h = IntPtr.Zero;
+            lock (this)
+            {
+                h = _handle;
+                _handle = IntPtr.Zero;
+            }
+            if (h != IntPtr.Zero)
+                c_api.TF_DeleteStatus(h);
+            GC.SuppressFinalize(this);
+        }
+
+        ~Status()
+        {
+            Dispose();
         }
     }
 }

src/TensorFlowNET.Core/TensorFlowNET.Core.csproj.DotSettings

@@ -0,0 +1,2 @@
+<wpf:ResourceDictionary xml:space="preserve" xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml" xmlns:s="clr-namespace:System;assembly=mscorlib" xmlns:ss="urn:shemas-jetbrains-com:settings-storage-xaml" xmlns:wpf="http://schemas.microsoft.com/winfx/2006/xaml/presentation">
+	<s:Boolean x:Key="/Default/CodeInspection/NamespaceProvider/NamespaceFoldersToSkip/=tensors/@EntryIndexedValue">True</s:Boolean></wpf:ResourceDictionary>

src/TensorFlowNET.Core/Tensors/Tensor.Creation.cs

@@ -31,7 +31,21 @@ namespace Tensorflow
         /// <summary>
         /// true if unmanaged buffer has been freed.
         /// </summary>
-        private bool deallocator_called => _deallocatorArgs.deallocator_called;
+        private bool _deallocator_called => _deallocatorArgs.deallocator_called;
+
+        /// <summary>
+        /// true if the Tensor was created from a managed array
+        /// </summary>
+        private bool _isPinnedArray => _deallocatorArgs.gc_handle != IntPtr.Zero;
+
+        /// <summary>
+        /// This holds values that are used by the unmanaged deallocator callback
+        /// </summary>
+        private DeallocatorArgs _deallocatorArgs = new DeallocatorArgs() { gc_handle = IntPtr.Zero };
+
+        // note: they must be assigned to a static variable in order to work as unmanaged callbacks
+        static Deallocator _hGlobalDeallocator = FreeHGlobalMemory;
+        static Deallocator _gcHandleDeallocator = FreeGCHandle;
 
         public Tensor(IntPtr handle)
         {
@@ -94,7 +108,7 @@ namespace Tensorflow
         {
             var v = (sbyte*)Marshal.AllocHGlobal(sizeof(sbyte));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(sbyte)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(sbyte), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(sbyte)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(sbyte), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -120,7 +134,7 @@ namespace Tensorflow
         {
             var v = (byte*)Marshal.AllocHGlobal(sizeof(byte));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(byte)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(byte), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(byte)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(byte), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -146,7 +160,7 @@ namespace Tensorflow
         {
             var v = (short*)Marshal.AllocHGlobal(sizeof(short));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(short)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(short), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(short)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(short), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -172,7 +186,7 @@ namespace Tensorflow
         {
             var v = (ushort*)Marshal.AllocHGlobal(sizeof(ushort));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(ushort)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(ushort), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(ushort)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(ushort), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -198,7 +212,7 @@ namespace Tensorflow
         {
             var v = (int*)Marshal.AllocHGlobal(sizeof(int));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(int)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(int), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(int)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(int), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -224,7 +238,7 @@ namespace Tensorflow
         {
             var v = (uint*)Marshal.AllocHGlobal(sizeof(uint));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(uint)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(uint), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(uint)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(uint), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -250,7 +264,7 @@ namespace Tensorflow
         {
             var v = (long*)Marshal.AllocHGlobal(sizeof(long));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(long)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(long), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(long)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(long), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -276,7 +290,7 @@ namespace Tensorflow
         {
             var v = (ulong*)Marshal.AllocHGlobal(sizeof(ulong));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(ulong)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(ulong), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(ulong)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(ulong), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -302,7 +316,7 @@ namespace Tensorflow
         {
             var v = (float*)Marshal.AllocHGlobal(sizeof(float));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(float)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(float), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(float)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(float), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -328,7 +342,7 @@ namespace Tensorflow
         {
             var v = (double*)Marshal.AllocHGlobal(sizeof(double));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(double)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(double), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(double)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(double), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
         
         /// <summary>
@@ -354,7 +368,7 @@ namespace Tensorflow
         {
             var v = (Complex*)Marshal.AllocHGlobal(sizeof(Complex));
             *v = value;
-            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(Complex)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(Complex), deallocator: hGlobalDeallocator, ref _deallocatorArgs);
+            _handle = TF_NewTensor(dType ?? dtypes.as_dtype(typeof(Complex)), dims:new long[0], num_dims: 0, data: (IntPtr)v, len: (UIntPtr)sizeof(Complex), deallocator: _hGlobalDeallocator, ref _deallocatorArgs);
         }
 #endif
 
@@ -444,7 +458,7 @@ namespace Tensorflow
                 dims.Length,
                 dotHandle,
                 (UIntPtr)buffersize,
-                hGlobalDeallocator,
+                _hGlobalDeallocator,
                 ref _deallocatorArgs);
 
             return tfHandle;
@@ -458,8 +472,7 @@ namespace Tensorflow
             _id = ops.uid();
         }
 
-        private bool _isPinnedArray => _deallocatorArgs.gc_handle != IntPtr.Zero;
-
+       
         /// <summary>
         /// Creates a new tensor from the given array without copying memory. The array is pinned down and the pointer passed on.
         /// </summary>
@@ -516,17 +529,11 @@ namespace Tensorflow
             var gcHandle = GCHandle.Alloc(data, GCHandleType.Pinned);
             _deallocatorArgs = new DeallocatorArgs() { gc_handle = GCHandle.ToIntPtr(gcHandle) };
             if (shape == null || shape.Length == 0)
-                return TF_NewTensor(dt, new long[0], 0, gcHandle.AddrOfPinnedObject() + start * element_size, (UIntPtr)(count * element_size), gcHandleDeallocator, ref _deallocatorArgs);
+                return TF_NewTensor(dt, new long[0], 0, gcHandle.AddrOfPinnedObject() + start * element_size, (UIntPtr)(count * element_size), _gcHandleDeallocator, ref _deallocatorArgs);
             else
-                return TF_NewTensor(dt, shape, shape.Length, gcHandle.AddrOfPinnedObject() + start * element_size, (UIntPtr)(count * element_size), gcHandleDeallocator, ref _deallocatorArgs);
+                return TF_NewTensor(dt, shape, shape.Length, gcHandle.AddrOfPinnedObject() + start * element_size, (UIntPtr)(count * element_size), _gcHandleDeallocator, ref _deallocatorArgs);
         }
 
-        private DeallocatorArgs _deallocatorArgs = new DeallocatorArgs() { gc_handle = IntPtr.Zero };
-
-        // note: they must be assigned to a static variable in order to work as unmanaged callbacks
-        static Deallocator hGlobalDeallocator = FreeHGlobalMemory;
-        static Deallocator gcHandleDeallocator = FreeGCHandle;
-
         [MonoPInvokeCallback(typeof(Deallocator))]
         internal static void FreeHGlobalMemory(IntPtr dataPtr, IntPtr len, ref DeallocatorArgs args)
         {

src/TensorFlowNET.Core/Tensors/Tensor.cs

@@ -350,12 +350,35 @@ namespace Tensorflow
 
         public void Dispose()
         {
-            if (_handle != IntPtr.Zero)
+            IntPtr h=IntPtr.Zero;
+            lock (this)
             {
-                c_api.TF_DeleteTensor(_handle);
-                _handle = IntPtr.Zero; 
+                h = _handle;
+                _handle=IntPtr.Zero;
             }
+            if (h != IntPtr.Zero)
+                c_api.TF_DeleteTensor(_handle);
             status.Dispose();
+            GC.SuppressFinalize(this);
+        }
+
+        /// <summary>
+        /// Dispose the tensor when it gets garbage collected
+        /// </summary>
+        ~Tensor()
+        {
+            Dispose();
+        }
+
+        public bool IsDisposed
+        {
+            get
+            {
+                lock (this)
+                {
+                    return _handle == IntPtr.Zero;
+                }
+            }
         }
 
     }

src/TensorFlowNet.Benchmarks/TensorCreation.cs

@@ -7,7 +7,7 @@ using Tensorflow;
 
 namespace TensorFlowNet.Benchmark
 {
-    [SimpleJob(launchCount: 1, warmupCount: 10, targetCount: 30)]
+    [SimpleJob(launchCount: 1, warmupCount: 2, targetCount: 10)]
     [MinColumn, MaxColumn, MeanColumn, MedianColumn]
     public class TensorCreation
     {
@@ -16,16 +16,19 @@ namespace TensorFlowNet.Benchmark
         [GlobalSetup]
         public void Setup()
         {
-            data = new double[1000];
+            data = new double[10];
         }
 
         [Benchmark]
         public void TensorFromArray()
         {
             var g=new Graph();
-            for (int i = 0; i < 100; i++)
+            for (int i = 0; i < 1000; i++)
             {
                 var tensor = new Tensor(data);
+                {
+
+                }
             }
         }
 
@@ -34,9 +37,12 @@ namespace TensorFlowNet.Benchmark
         public void TensorFromNDArray()
         {
             var g = new Graph();
-            for (int i = 0; i < 100; i++)
+            for (int i = 0; i < 1000; i++)
             {
                 var tensor = new Tensor(new NDArray(data));
+                {
+
+                }
             }
         }
 

test/TensorFlowNET.UnitTest/TensorTest.cs

@@ -5,17 +5,54 @@ using System.Collections.Generic;
 using System.Linq;
 using System.Runtime.InteropServices;
 using System.Text;
+using System.Threading;
+using System.Threading.Tasks;
 using Tensorflow;
+using static Tensorflow.Python;
 
 namespace TensorFlowNET.UnitTest
 {
     [TestClass]
     public class TensorTest : CApiTest
     {
-        /// <summary>
-        /// Port from c_api_test.cc
-        /// `TEST(CAPI, AllocateTensor)`
-        /// </summary>
+        [TestMethod]
+        public void TensorDeallocationThreadSafety()
+        {
+            var tensors = new Tensor[1000];
+            foreach (var i in range(1000))
+            {
+                tensors[i] = new Tensor(new int[1000]);
+            }
+            SemaphoreSlim s = new SemaphoreSlim(0, 2);
+            SemaphoreSlim s_done = new SemaphoreSlim(0, 2);
+
+            var t1 = new Thread(() =>
+            {
+                s.Wait();
+                foreach (var t in tensors)
+                    t.Dispose();
+                s_done.Release();
+            });
+
+            var t2 = new Thread(() =>
+            {
+                s.Wait();
+                foreach (var t in tensors)
+                    t.Dispose();
+                s_done.Release();
+            });
+
+            t1.Start();
+            t2.Start();
+            s.Release(2);
+            s_done.Wait();
+            s_done.Wait();
+
+            foreach(var t in tensors)
+                Assert.IsTrue(t.IsDisposed);
+        }
+
+
         [TestMethod]
         public void AllocateTensor()
         {
@@ -29,6 +66,7 @@ namespace TensorFlowNET.UnitTest
             t.Dispose();*/
         }
 
+
         /// <summary>
         /// Port from c_api_test.cc
         /// `TEST(CAPI, MaybeMove)`