Date: Thu, 5 Jun 2003 21:08:38 -0600 From: <soralx@cydem.org.ua> To: peterjeremy@optushome.com.au, zec@tel.fer.hr Cc: hackers@freebsd.org Subject: Re: Network stack cloning / virtualization patches Message-ID: <200306052108.38256.soralx@cydem.org.ua> In-Reply-To: <0HFZ00KBVWQTKQ@l-daemon> References: <0HFZ00KBVWQTKQ@l-daemon>
next in thread | previous in thread | raw e-mail | index | archive | help
> A third issue on the x86 is a lack of registers: There are only 6 > "general purpose" registers (and each of them actually has a specific > purpose). Eating one of these registers to maintain a pointer to > a struct vimage will be a noticable performance hit. Why not to store it in memory? If the pointer needs to be read often, there is a high chance that it will be in CPU's data cache. Nowadays with these huge FSB frequencies reading dword from RAM won't be _much_ slower that reading it from a register anyway. IMO, experiment is req'd to see if you will win some performance be freeing a register. Or am I wrong? 03.06.2003; 21:30:43 [SorAlx] http://cydem.org.ua/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200306052108.38256.soralx>