CLFORTRAN is a new and elegant Fortran module that allows integration of OpenCL with Fortran programs easier than ever.
Taking advantage of Fortran language features, it is written in pure Fortran – aka no C/C++ code is required to utilize the GPGPU.
CLFORTRAN is compatible with all major compilers: GNU, Intel and IBM, and supporting OpenCL 1.2 API.
In addition, it is provided as open source and licensed under LGPL, to allow scientific computing at massive scales and all supported vendors.
Intel® have just introduced their Xeon® Phi™ processor to the market, targeting HPC and scientific computing. It is available for purchase and integration into existing systems/platforms/servers.
The new co-processor is a discrete device that runs an operating system of its own and functions as a fully functional computer (though being a co-processor).
Why should anyone be interested in Xeon® Phi™?
It is based on the most common x86 architecture, therefore porting existing code and algorithms should be the easiest possible.
One may also utilize OpenCL™ algorithms to take advantage of the high-parallelism of the Phi™ processor.
In addition, it features 60 cores, 8GB of internal memory (with 320 GB/s) and uses PCIe x16 slot to provide high performance bandwidth throughput.
With almost 1 TFLOPS of double precision, Phi™ is competent to very high end GPUs in the market today, but on some aspects, provides better performance and industrial matching than other vendors.
We are happy to announce the release of CUDA.NET version 3.0.0.
This release provides support for latest CUDA 3.0 API and few more updates that will make programming with CUDA from .NET easier and faster.
Support for CUDA 3.0 API
Added memset functions for CUDA class
Supporting new graphics interoperability functions
Improved generics support for memory operations
Added CUDAContextSynchronizer class
Improved memory operations
We employ GCHandle class to be used with generic memory copies in CUDA class. This method allows to work with every data type (existing vectors or user defined) natively in .NET. The implication is that now you can copy existing custom arrays of structures/classes (user data-types) to device with memory copy functions.
This class was added to assist developers in multi-GPU and multi-threaded environments sharing the same device. It uses existing CUDA API to manipulate the context each thread is attached to and provides .NET means to synchronize between threads sharing the same device for different computations.
Find it under the Tools namespace, the documentation includes a description of how to use it.
We hope you will enjoy this release.
As always, please send us comments or suggestions to: firstname.lastname@example.org.
We are happy to announce the availability of the so long waiting OpenCL.NET 1.0.48 library.
This version aligns with OpenCL 1.0.48 standard, and fully conforms with latest NVIDIA drivers for OpenCL (and as well on supported platforms).
In brief, this release of the standard added few API functions and modified some, to truly allow heterogeneous computing on a single system. An application can query for the existence of multiple computing devices on the system, also by different vendors (recognize the CPU and a GPU as compute resources) regardless of the vendor. Such that consuming different computing resources can be transparent.
For further details about standard features and changes please consult Khronos website.