LWN: Comments on "Introducing Fedora Nightlife (Bryan's Blog)" https://lwn.net/Articles/284448/ This is a special feed containing comments posted to the individual LWN article titled "Introducing Fedora Nightlife (Bryan's Blog)". en-us Tue, 28 Oct 2025 16:13:10 +0000 Tue, 28 Oct 2025 16:13:10 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net Introducing Fedora Nightlife (Bryan's Blog) https://lwn.net/Articles/284486/ https://lwn.net/Articles/284486/ jmorris42 <div class="FormattedComment"><pre> <font class="QuotedText">&gt; There's a significant power draw associated with actually using the CPU...</font> Yup, this idea is so 1990's. Back when the SETI@Home and various early crypto challenges were running most of the clients were Windows screensavers running on machines that didn't even do idle instructions and you were lucky to find a laptop that would reliably sleep or suspend. In that sort of setting cycles really were going to waste and recovering them for a productive purpose made sense because patching the OS wasn't really possible. But again, you nailed it. Nowadays Windows does better power manegement than even we do, although we are catching up again. Ramping a CPU (and the fans) up all night to run some cpu intensive yet likely dubious 'scientific' project makes less sense. </pre></div> Fri, 30 May 2008 03:14:50 +0000 Not just Fedora https://lwn.net/Articles/284484/ https://lwn.net/Articles/284484/ midg3t <p>I expect that it will be renamed once users of other distros get involved.</p> Fri, 30 May 2008 02:26:50 +0000 Introducing Fedora Nightlife (Bryan's Blog) https://lwn.net/Articles/284463/ https://lwn.net/Articles/284463/ johnkarp <div class="FormattedComment"><pre> Maybe some of the volunteers do in fact know what they're doing. For a small scientific projects, its probably much easier to accept donations of CPU time, than it is to solicit funds/grants and run a compute farm. </pre></div> Thu, 29 May 2008 22:31:41 +0000 Introducing Fedora Nightlife (Bryan's Blog) https://lwn.net/Articles/284456/ https://lwn.net/Articles/284456/ mjg59 <div class="FormattedComment"><pre> The very existence of idle capacity is a failure (idle machines should be suspended or turned off), but a more serious one is the implicit assumption that tends to be made in these situations - that is, that the marginal cost of running jobs on otherwise idle hardware is zero. It's not. There's a significant power draw associated with actually using the CPU, to the extent that I've been told that the difference in draw between using a tickless kernel and a 100Hz kernel on some modern machines can be on the order of 20 Watts when idle. As we push more power saving functionality into the OS and the hardware that number is only going to get larger. Large-scale computation belongs in centralised facilities where economies of scale mean you can drop the overall power consumption, not spread across a large number of well-meaning volnteers who have no idea what it's costing them. </pre></div> Thu, 29 May 2008 21:33:53 +0000