Psi Lambda LLC | Parallel computing made practical.

The NVIDIA Driver version 256.XX or higher is needed for all CUDA 3.1software–including Kappa version 1.3 or higher. The NVIDIA Driver version 195.36.15 or higher is needed for all CUDA 3.0 software–including Kappa versions 1.0 – 1.2. You will need the appropriate rpm files from the Psi Lambda Download/Kappa for Linux. Import Psi Lambda LLC [...]

· · · ·

The NVIDIA Driver version 256.XX or higher is needed for all CUDA 3.1software–including Kappa version 1.3 or higher. The NVIDIA Driver version 195.36.15 or higher is needed for all CUDA 3.0 software–including Kappa versions 1.0 – 1.2. You will need the appropriate rpm files from the Psi Lambda Download/Kappa for Linux. Import Psi Lambda LLC [...]

· ·

Kappa can show an easy ten times speed increase on GF100 GPUs (compared to CUDA driver or runtime 3.1 API programs) Here are two Kappa scheduling scripts that only differ by the assignment of a stream to a kernel–one has the same stream assigned to all kernel executions so that the kernels execute sequentially and [...]

· · ·

May/10

28

Kappa Extras–NVIDIA CUDA coming to a language near you

The Kappa Extras are available for download. These MIT licensed extras allow using CUDA via the Kappa Library from many more programming languages. These are now (as of Kappa version 1.2) distributed in the Kappa installation. Extras for Kappa via SWIG The following files allow wrapping the Kappa Library for use by the languages supported [...]

· · · · · · · · · · · · · · ·

The Kappa Library makes creating CUDA based applications and libraries easy and flexible. While NVIDIA gives you the tools to develop CUDA kernels–Kappa gives you a better ability to coordinate and integrate those kernels as part of an application or library. You can even use the same ‘.cu’ files without any changes (for CUDA kernel [...]

No tags

May/10

24

Kappa for Python is now available

Kappa now makes using CUDA from Python even easier–with Kappa’s memory management, kernel scheduling and other features. The KappaCUDA package is now available.

·

May/10

24

Kappa for Perl is now available

The ability to use CUDA from Perl is now available–see the download page for Kappa for Perl for details. The keyword Perl and the KappaCUDA module is now available. The description is available on CPAN also.

· ·

It is possible to do a simplistic performance comparison for the Kappa Library to give a rough idea of how the Kappa Library will perform relative to programs or libraries written directly with the CUDA driver or runtime API’s. The results given here are only meant to provide a starting point for comparison until you [...]

·

<< Latest posts

Articles