EM Photonics' CULA linear algebra library brings GPU computing to developers

SANTA CLARA, USA: EM Photonics released a beta version of CULA, an implementation of the industry-standard LAPACK linear algebra library designed and optimized for NVIDIA’s massively parallel CUDA-enabled graphics processing units (GPUs).

The millions of developers that rely on LAPACK routines for solving problems ranging from computational physics and structural mechanics to electronic design automation can now get up to a 10X boost in performance over a single quad-core CPU1 by using NVIDIA Tesla GPUs in their workstation or datacenter.

"One promising evolutionary path of high-performance computing architectures is a hybrid system consisting of multi-core CPUs and many core GPUs," said Professor Satoshi Matsuoka, of the Tokyo Institute of Technology.

"LAPACK is key for many scientific applications, so a CUDA-optimized implementation will significantly broaden the appeal of hybrid systems in science and engineering, giving them a strong competitive edge over competing architectures"

“We began a partnership with NASA Ames Research Center to create GPU-accelerated linear algebra libraries in 2007,” said Eric Kelmelis, CEO of EM Photonics. “As an offshoot of this project and through a partnership with NVIDIA, EM Photonics is releasing CULA and allowing developers to experience the computational performance of a supercomputer right at their desk.”

EM Photonics’ CULAtools is a product family comprised of CULA Basic, Premium, and Commercial. The CULA library is a GPU-accelerated implementation of the most popular LAPACK routines.

LAPACK is a collection of commonly used functions in linear algebra, used by millions of developers in the scientific and engineering community. The problems they tackle can often be approximated by linear models and can, therefore, be solved using linear algebra routines. CULA exploits the massively parallel CUDA architecture of NVIDIA’s GPUs to accelerate many of the common LAPACK routines.

“Our customer base has been anticipating the release of a linear algebra library similar to LAPACK. This fundamental math library brings the power of GPU computing to a much broader developer base in the scientific computing community”, said Andy Keane, general manager of the Tesla business unit at NVIDIA.

“CULA forms yet another key branch in our rapidly increasing ecosystem of CUDA libraries which now includes FFT, BLAS, image processing, computer vision, ray tracing, rendering, molecular dynamics, and more.”

A full production release of CULA is scheduled for NVIDIA’s GPU Technology Conference, being held from Sept. 30-October 2nd at the Fairmont Hotel in San Jose, California.

PGI, NVIDIA to deliver CUDA Fortran compiler

SINGAPORE: The Portland Group, a wholly-owned subsidiary of STMicroelectronics and a leading supplier of compilers for high-performance computing (HPC), today announced an agreement with NVIDIA under which the two companies plan to develop new Fortran language support for CUDA GPUs.

The NVIDIA CUDA architecture allows developers to offload computationally intensive kernels to the massively parallel GPU. Through function calls and language extensions, CUDA gives developers explicit control over the mapping of general-purpose computational kernels to GPUs as well as placement and movement of data between the x64 processor and the GPU.

The NVIDIA CUDA C compiler already provides this capability to C programmers. The CUDA Fortran compiler will provide this same level of control and optimization in a native Fortran environment from PGI.

“Fortran support for CUDA GPUs is a perfect complement to our existing roadmap for the PGI Accelerator Fortran and C compilers,” said Douglas Miles, director, The Portland Group. “It enables interoperability of PGI Fortran and CUDA C and gives PGI users a full range of options in porting and optimizing Fortran applications to leverage the power of CUDA-enabled NVIDIA GPUs.”

“The GPU computing developer community has made it clear there is a need and demand for a production-quality Fortran solution on the GPU,” said Andy Keane, general manager, Tesla GPU Computing Solutions, NVIDIA. “With their large base of Fortran developers for x64 processor-based HPC systems, PGI provides a perfect bridge for migration of production science and engineering codes from existing platforms to NVIDIA Tesla GPUs.”

The Portland Group and NVIDIA will release the Fortran language specification for CUDA GPUs at the International Conference on Supercomputing in Hamburg, Germany this week. The CUDA Fortran compiler will be added to a production release of the PGI Fortran compilers scheduled for availability in November 2009.