“Auto-magic” GPU Programming

We all agree, GPUs are not as easy to program as a CPU, even though CUDA and OpenCL provide a well established abstraction of the GPU processor design. GPU support in high-level programming languages such as Java, C# or Python still leaves a lot to be desired. Only developers familiar with C++ can directly access a rich set of APIs and tools for general-purpose GPU programming.

We believe this does not need to be so. This blog is a sneak preview of the new upcoming version 3 of Alea GPU which sets completely new standards for GPU development on managed platforms such as .NET or the JVM. So what is so exciting about version 3?

  1. It adds higher level abstractions such as GPU LINQ, a GPU parallel-for and parallel aggregate, which can build and execute GPU workflows or run delegates and lambda expressions on the the GPU. An additional benefit is code reuse – the same delegate can be executed on a GPU as well as on the CPU.

  2. It makes GPU programming much easier and often “auto-magic” – we manage memory allocation and transfer to and from the GPU in an economic way. It is built for developers who grew up without pointers, malloc and free.

  3. We integrate well with the .NET type system, which means for example that .NET arrays can be used directly in GPU kernels.

  4. Of course for all the hard core CUDA programmers with .NET affinity, we still support CUDA in C# and F# as we did already with our previous versions – even better.

Let’s look at the API from a usability point of view and at the underlying technology and implementation details.


LINQ is a technology that extends languages such as C# with powerful query capabilities for data access and transformation. It can be extended to support virtually any kind of data store, also data and collections that reside on a GPU. Alea GPU LINQ introduces new LINQ extensions to express GPU computations with LINQ expressions that are optimized and compiled to efficient GPU code. The main advantages of coding whole GPU workflows with LINQ expressions are:

  1. The execution of GPU LINQ workflows is delayed. They are composable , which facilitates code modularity and reuse.

  2. GPU LINQ workflows provide many standard operations such as parallel aggregation or parallel map, which makes them more expressive and reduces boiler plate code.

  3. We can apply various kernel optimization techniques, such as GPU kernel fusing, which results in fewer GPU kernel launches and more compact GPU code.

  4. Having the full code as an expression allows us to better optimize memory management and data transfer.

Here is a simple but interesting example, which determines the largest value of a array of values on the GPU in two steps. First we index the sequence and then we reduce the new array of indexed values with a custom binary operator that just compares the values to find the maximal value and its index. A priori this would require two GPU kernels. The first is a parallel map, the second is a parallel reduction. Alea GPU can fuse the two kernels into one.

A more complex example is the calculation of the fair value of an Asian option with Monte Carlo simulation based on the celebrated Black-Scholes model.

The Monte Carlo simulation runs in multiple batches, each batch consists of numSamplesPerBatch samples. Workflows are composable. The outer workflow is a map-reduce, which launches the batch sampling and reduces the match mean to the mean across all batches. The inner workflow does the actual Monte Carlo simulation. It first binds storage to the workflow which is then populated with normally distributed random numbers. The core algorithm is in the Select. For each sample index i it generates a path and prices the option along the path. The Aggregate method of the inner workflow calculates the batch sample mean with a parallel reduction.

Black Scholes Simulation

GPU Parallel-For and Parallel Aggregate

An alternative abstraction is provided with the GPU parallel-for and parallel aggregate pattern. Together with automatic memory management, they allows us to write parallel GPU code as if you would write serial CPU code. The usability is very simple. We select a GPU device and pass a delegate to the gpu.For method. All the variables used in the delegate are captured in a closure that is then passed to the parallel-for body. The data is automatically moved to the GPU and the results are brought back automatically to the host.

Parallel-For Closure

The element-wise sum on the GPU is now as simple as this:

The delegate accesses data elements arg1 and arg2 that are defined outside of the loop body and writes the result directly to a .NET array. The runtime system takes care of all the memory management and transfer. Because the delegate does not rely on any GPU specific features such as shared memory,the delegate can execute on the CPU and the GPU. The runtime system also takes care of selecting the thread block size based on the occupancy of the generated kernel.

The parallel aggregate works in the same way. It requires a binary associative operator which is used to reduce the input collection to a single value. Our implementation does not require that the operator is commutative. The following code calculates the sum of the array elements on the GPU:

Automatic Memory Management

Our automatic memory management system handles memory allocation and data movement between the different memory spaces of the CPU and the GPU without the programmer having to manage this manually. It is efficient – unnecessary copy operations are avoided by analyzing the memory access. The implementation is based on code instrumentation, a technique that inserts additional instructions into an existing execution path. Alea GPU modifies the CPU code by inserting instructions that monitor array accesses and perform minimum data transfers between the CPU and GPU. As these runtime checks generate a slight performance overhead, the scope of analysis is limited to the code carrying the attribute [GpuManaged]. Leaving out this attribute never means that data will not be copying – it may only affect unnecessary intermediate copying.

To illustrate the automatic memory management in more detail, we look at an example. We iterate 100 times a parallel-for loop that increments the input by one. First of all, we consider the situation without the [GpuManaged] attribute. In this case, the data is automatically copying, although more frequently than necessary due to a limited scope of analysis.

We check the memory copy operations by using the NVIDIA Visual Nsight profiler. As expected the low level CUDA driver functions cuLaunchKernel, cuMemcpyHtoD_v2 and cuMemcpyDtoH_v2 to launch the kernel and to perform memory copy are called 100 times each. This means that the data is copied in and out for each of the 100 sequential parallel for launches. Let us add the attribute [GpuManaged] to turn on automatic memory management.

We see that cuMemcpyHtoD_v2 and cuMemcpyDtoH_v2 are now called just once. The reason is that result data of a preceding GPU parallel-for loop can stay on the GPU for the succeeding parallel-for loop without need of copying the intermediate data back and forth to CPU. Copying is only involved for the input of the first GPU execution as well as for the output of the last GPU computation.

Using .NET Arrays and Value Types in Kernels

For a C# developer it would be very convenient to use .NET arrays and other standard .NET types also directly in a GPU kernel and that all the memory management and data movement is handled automatically. .NET types are either reference types or value types. Value types are types that hold both data and memory at the same location, a reference type has a pointer which points to the memory location. Structs are value types and classes are reference types. Blittable types are types that have a common representation in both managed and unmanaged memory, in particular reference types are always non-blittable. Copying non-blittable types from one memory space to another requires marshalling, which is usually slow.

From the point of view of efficiency we made the decision to only support .NET arrays with blittable element types as well as jagged arrays thereof. This is a good compromise between usability and performance. To illustrate the benefits let’s look at how to write an optimized matrix transpose. With Alea GPU version 2 you have to work with device pointers and all the matrix index calculations have to be done by hand.

Alea GPU version 2 requires that kernels and other GPU resources are in a class that inherits from ILGPUModule. Apart from this the kernel implementation resembles the CUDA C implementation very closely.

With Alea GPU V3 you don’t need to inherit from a base module class anymore. You can directly work with .NET arrays in the kernel, also for the shared memory tile. We save the error prone matrix element index calculations and only need to map the thread block to the matrix tile.

Alea GPU version 2 requires explicit memory allocation, data copying and calling the kernel with device pointers. An additional inconvenience is that matrices stored in two-dimensional arrays first have to be flatten.

Here is the kernel launch code that relies on automatic memory management. The developer allocates a .NET array for the result, passes that, together with the input matrix directly to the kernel.

Without compromizing the usability, the programmer can also work with explicit memory management.

Here the arrays a and at are fake arrays representing arrays on the GPU device and he can use them in a GPU kernel the same way as an ordinary .NET array. The only difference is that he is now responsible to copy back the result explicitly with CopyToHost. Of course the deviceptr<T> API is still available and often useful for low level primitives or to write highly optimized code.

Improved Support for Delegates and Generic Structs

Alea GPU version 3 also has better support for delegates and lambda expressions. Here is a simple generic transform that takes a binary function object as argument and applies it to arrays of input data:

We can launch it with a lambda expression as follows:

The next example defines a struct representing a complex number which becomes a blittable value type.

We define a delegate that adds two complex numbers. It creates the result directly with the default constructor. Note that this delegate is free of any GPU specific code and can be executed on the CPU and GPU alike.

It can be used in the parallel Gpu.For to perform element-wise complex addition

or in the above generic transform kernel.

JIT Compiling Delegates to GPU Code

From an implementation point of view a challenge is that delegates are runtime objects. This means we have to JIT compile the delegate code at runtime. Fortunately our compiler has this feature since its initial version. For a delegate such as

the C# compiler will generate a closure class with fields and an Invoke method:

To instantiate the delegate instance, the C# compiler generates code to instantiate the closure class, set its fields, and to create the delegate instance with the closure instance and the method’s function pointer:

The above code is just illustrative and not legal C# code. Both ldftn and methodof are in fact the
real IL instructions that C# compiler generates.

Whenever the Alea GPU compiler finds this delegate, it translates the closure class into a kernel struct, and JIT compiles the GPU kernel code that comes from the Invoke method of the compiler generated class. Alea GPU caches the result of JIT compilations in a dictionary using the key methodof(CompilerGenerated.Invoke), so it will not compile delegates with same method multiple times.

There is one thing needs to be noted. Since we translate the closure class into a struct and pass it to the GPU as a kernel argument, it is not possible to change the values of the fields. For example a delegate like i => result = arg1 does not work.

Code Instrumentation for Automatic Memory Management

The core component of our automatic memory management system is a memory tracker. It tracks .NET arrays and their counterparts residing in GPU device memory. Every array has a flag that indicates if it is out of date. The tracking of an array starts the first time it is used (implicitly in a closure or explicitly as an argument) in a GPU kernel launch. A weak reference table stores for every tracked array the host-out-of-date flag and for every GPU the corresponding device memory, together with the device-out-of-date flag.

The memory tracker has the following methods:

  1. Start tracking a host array
  2. Make an array up to date on a specific GPU
  3. Make an array up to date on host
  4. Make all arrays up to date on a specific GPU
  5. Make all arrays up to date on host

The default procedure is as follows: If an array is used in a kernel launch on a GPU the tracker makes the array up to date on that GPU by copying it to device memory just before the kernel is launched. After the kernel launch the tracker makes the array again up to date on the host by copying it back to the CPU memory. This very simple strategy always works but often leads to unnecessary memory transfers. The basic idea of our automatic memory management system is to defer the synchronization of a host arrays with its device counterpart to the point when the host array is actually accessed again. We implement this deferred synchronization strategy with code instrumentation, which inserts additional checks and memory tracker method calls at the right place.

Because instrumentation adds additional overhead we narrow down the ranges of instrumentation. A function can either be GpuManaged or GpuUnmanaged. By default, a function is GpuUnmanaged, which means that it does not defer the memory synchronization and thus its code is not instrumented. If a function has the GpuManaged attribute, we insert code and method calls to track the array access and defer the synchronization. At least, the functions Alea.Gpu.Launch and Alea.Gpu.For are GpuManaged.

Methods with the attribute GpuManaged are inspected in a post-build process. We check if a function contains IL instructions such as ldelem , ldelema , stelem, call Array.GetItem(), call Array.SetItem(), etc. to access a specific array. In this case we extract the array operand and insert code to defer its synchronization. A standard use case is a loop over all the array elements to set or modify them. In such a case we can optimize the tracking by creating local cache flags. Here is an example:

Instrumention produces code that is functionally equivalent to the following source code:

Calling a method like MemoryTracker.HostUpToDateFor() many times to check if an array has to be synchronized is generating a huge overhead. We use the flag to bypass the call once we know the array is synchronized and reset the flag after kernel launches. At the end of GpuManaged method, we will insert code to bring all out-of-date implicitly traced array back to host. A frequent case is calling other functions from a GpuManaged function. These other functions could be either GpuManaged or GpuUnmanaged. We need to notify the callee to defer memory synchronization. We use some mechanism to pass the managed session to the callee, so that it won’t bring back all out-of-date array to host, because it is not the end of the GpuManaged session.

The implementation relies on Mono.Cecil and Fody. There is a sketch of the full code instrumentation that is executed in a post build step:

  1. Load the compiled assemblies with Mono.Cecil through Fody
  2. For each GpuManaged function
    1. Add memory barrier code
      • for every array element access add cache flag and call to HostUpToDateFor()
      • for GpuManaged functions call SetFlagOnCurrentThread() before, reset all cache flags after
      • for GpuUnmanaged functions call HostUpToDateForAllArrays() before calling them
    2. Add try finally block and in finally call HostUpToDateForAll() if the caller is GpuUnmanaged
  3. Weave the modified assembly via Fody

Your Feedback

We hope that after reading this post you share the same excitement for the new upcoming version 3 of the Alea GPU compiler for .NET as we do.

Of course we are interested to hear all of your feedback and suggestions for Alea GPU. Write to us at or @QuantAlea on Twitter.

The features that we presented here are still in preview and might slightly change until we finally release version 3.

If you would like to already play around with Alea GPU V3 come and join us April 4 at GTC 2016 and attend our tutorial on simplified GPU programming with C#.