Date: Wed, 22 Mar 2000 17:58:58 -0800 From: "Jeremiah Gowdy" <jgowdy@home.com> To: "James" <jkelty@digital-impact.com>, <freebsd-questions@FreeBSD.ORG> Subject: Re: SMP specifics.......????????? Message-ID: <000f01bf946b$5ae2fca0$0100000a@vista1.sdca.home.com> References: <CPEEKAEIEILAHOOFPBOICEDCCDAA.jkelty@digital-impact.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> Does the SMP support allow a process running on each CPU to talk to a single > device with locking the kernel up? It should act the same way as if two processes accessed the device at the same time. Shouldn't lock anything up. >Is it true SMP, or just a little more > processing power? How could it not be true SMP ? What would qualify as just a little more processing power ? The only way I could see it being implemented, on any system, not just FreeBSD, is that the task scheduler runs two processes at once, one on each processor. That's the only way a second processor would add any more processing power. A priority queue with two places for output/processing rather than one. I can't really think of a cheezy way to do it that wouldn't qualify as SMP. Now, I know SMP has other features, like processor affinity and whatnot, of which I'm not familiar, which would be a better implementation. However, even the crappiest implementation of SMP is still "true SMP". As for FreeBSD's SMP, I've not seen any specifics on the implemenatation, except that I hear 4.0 includes linux/processor threads, which allows a multithreaded program's threads to use either processor, allowing a single highly active multithreaded process to use the SMP setup. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?000f01bf946b$5ae2fca0$0100000a>