Date: Wed, 4 Jun 2008 21:35:00 -0700 (PDT) From: Unga <unga888@yahoo.com> To: freebsd-chat@freebsd.org Subject: Supercomputing with FreeBSD? Message-ID: <285746.27928.qm@web57009.mail.re3.yahoo.com>
next in thread | raw e-mail | index | archive | help
Hi all supercomputing interested guys and gals You may have seen this: 1. University of Antwerp makes 4000EUR NVIDIA supercomputer (http://www.dvhardware.net/article27538.html) 2. FASTRA GPU SuperPC (http://fastra.ua.ac.be/en/index.html) I would like to first quote following from http://fastra.ua.ac.be/en/specs.html : "Software overview ------------------ We selected Windows XP-64 as the operating system for FASTRA. There were three reasons for choosing this platform: first, we needed a 64-bit operating system, in order to utilize 8GB of RAM. Second, we expected fewer driver issues on Windows compared to Linux. Third, within the Windows product line, Windows Vista is not yet supported by the NVIDIA GPU Computing platform, leaving Windows XP as the only choice. For development, we use Microsoft Visual Studio 2008. The core functionality for our CPU code is written in C++ (Visual C++), while MATLAB is often used as a front-end for rapid prototyping. All GPU code is developed using the NVIDIA CUDA framework (http://www.nvidia.com/object/cuda_home.html), a C-like programming language that allows for efficient programming of the NVIDIA GPUs." This opportunity make it available to FreeBSD users has many great benefits. We can use FreeBSD, AMD64 and Nvidia combination at an affordable price for great many computational intensive tasks such as compilation (FreeBSD has a parallel make), rendering, encoding, etc. Of course such supercomputational-ready software should be available first. But the question is, is the FreeBSD infrastructurally ready for that? FreeBSD runs on amd64. But we have following issues: 1. Nvidia doesn't release a driver for amd64. 2. The NVIDIA CUDA framework (http://www.nvidia.com/object/cuda_home.html), is not available for FreeBSD, but it is available for Linux and Mac OSX. So porting CUDA to FreeBSD may not be a big issue. To resolve the above two issues: 1. FreeBSD should proactively address following issues: - http://lists.freebsd.org/pipermail/freebsd-hackers/2006-June/016995.html - http://wiki.freebsd.org/NvidiaFeatureRequests I don't understand why the the FreeBSD project does not organize a Google SoC style project to address the above issues, invite few developers to join the project, either use FreeBSD donated funds or seeks fresh funds for the project (AMD and Nvidia will sure donate if requested as they are direct beneficiaries). The project could be at least to be targeted to commit for upcoming FreeBSD 8.0. I would like to understand why organize such a project is very difficult and what are the issues regarding that. 2. Once above point 1. is fixed, I'm sure the Nvidia will port the CUDA framework to FreeBSD and release a driver for amd64. If not FreeBSD project/foundation can request from Nvidia. Kind regards Unga
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?285746.27928.qm>