I’ve been a fan of distributed computing since late 1990s, with SETI@Home running on every computer that I ever had. However, the real attractive proposition to me was running distributed computing applications on graphics cards. GPUs are much more efficient in stream computing than any CPU you could find, and I’ve tried DC apps on computers with DEC Alpha, Intel Pentium onwards, AMD K6-II onwards etc etc., but biggest jump in performance was Folding@Home on ATI Radeon X1800XTX graphics card.
With the launch of this blog and the new website, I’ve decided to launch a new group, number 69864. Current name is the name of this blog, but as soon as I am able to disclose the name of the new company, you’ll be the first to know 😉
I invite you all to join team 69864, and in near future this will be much more than just a group. As the new website develops, so will this team. So far, my goal is to enter Top 1000 by end of 2008, and we’re on good way to achieve that.
If you are interested, the site to download CPU clients is here, while the GPU clients are here. If you have Nvidia-based graphics card (compatible from GeForce 8000 and above), then download this version. I’ve tried both console and the regular version, and there isn’t much difference in performance. Of course, unless you leave the display version running.
Performance-wise, Nvidia is destroying ATI at this moment, which is something I already addressed here. Let’s hope ATI will optimize Folding performance in their upcoming drivers. We doubt this will happen before the release of next generation of Radeon hardware, but who knows.
After you install the client, the configuration is really easy. If you have higher-performing hardware, then always use the large packet option, as shown in picture below.
Initial option screen... put your name in, and the group is 69864 😉
If you have 256MB or more video memory, check this box
MachineID is important if you plan to run more than one client (multi-GPU setups, CPU/GPU combo etc.)
After configuration, you’re good to go. The stats page is located at KakaoStats.com – yes, Kakao means Cocoa in Croatian ;-). You can also see the official one here, but the official stats page isn’t always available, since F@H servers suffer from tremendous load.
Let’s fold together and hopefully simulate enough nanoseconds that become seconds that become minutes, hours, days, years… who knows, our CPU or GPU time might help scientists tofind cure for Alzheimer or Parkison’s disease (focus of F@H group at Stanford). We might even help ourselves in the future.
Unofficial stats over at KakaoStats.com - enjoy this one even if you don't use our group #.
Pay it forward. You’re free to ping me at theo.valich @ gmail.com for a chat or if you have any questions. Always glad to help 😉
