FASTEST WAY TO PORTABLE COMPUTING
Ideal for Accelerating Legacy Codes
OpenACC directives are the fast, simple and portable way to accelerate your scientific code. With OpenACC, you insert compiler hints – in the form of OpenMP-like directives – into the compute-intensive portions of even the largest, most complex Fortran or C application, and the compiler automatically maps that code to an accelerator – including NVIDIA GPUs -- for higher performance. (OpenACC is fully compatible — and interoperates — with OpenMP and MPI.)
Now available from industry leaders - Cray, CAPS entreprise, and The Portland Group (PGI), OpenACC compilers are:
- Portable: Future-proof your codes with this open standard
- Fast: Straight forward, high-level, compiler driven approach to parallel computing
- Powerful: Ideal for accelerating large, legacy Fortran or C codes
Here's an example of how a single directive hint can accelerate the calculation of pi. With accelerator directives, you can get started and see results in the same afternoon.
Try it yourself with a free 30-day trial of the PGI Accelerator Compiler.
* Registration requires creating an account on the PGI website. After registering, you'll receive an email from PGI with further details.
"The PGI compiler is now showing us just how powerful it is. On the software we are writing, it's many times faster on the NVIDIA card. We are very pleased and excited about the future uses. It's like owning a personal supercomputer."
Dr. Kerry Black
University of Melbourne
"I have written micromagnetic codes (written in Fortran 90) to study the properties of two and three dimensional magnetic systems. The directives approach enabled me to port my existing code with ease to perform my computations on the GPU which resulted in a significant speedup of the computation."
Professor M. Amin Kayali
University of Houston
Read more success stories...
integer, parameter :: n=1000000
integer :: i
real(kind=8) :: t, pi
pi = 0.0
!$acc parallel loop
do i=0, n-1
t = (i+0.5)/n
pi = pi + 4.0/(1.0 + t*t)
!$acc end parallel loop
print *, 'pi=', pi/n
end program picalc
2X in 5 Steps
Most developers who try OpenACC see speedups of from 2 to 10X, following five key steps:
- Evaluate and plan
- Add parallel directives
- Add data movement directives
- Tune data movement
- Optimize parallel scheduling
To learn more about OpenACC Directives:
- Watch the 5-minute CUDAcast
- Watch the video “Getting Started with OpenACC” (13 minutes)
- Learn more about 2X in 5 Steps. Visit the Getting Started page at www.openacc.org
You'll need an NVIDIA® CUDA-capable GPU and a Fortran or C based application to get started. Don't have an accelerator? We recommend an NVIDIA Tesla K20 GPU.