Nathan Coulson wrote: > I was wondering, if using --enable-kvm --drive file=os.img,if=virtio > would give any noticable improvement > > --enable-kvm, use the virtualization instructions in the processor (faster) > --drive file=os.img,if=virtio, (Requires VIRTIO_BLK in kernel), > > threw a thread up on blfs-dev's virtualization topic that flushed out > the above a bit
I'll give that a try. > Could be worth checking the difference between using the graphical > output, as opposed to ssh as well. In the past, I've noticed some > slow compilations due to the computer trying to scroll the text. > (xorg, intel's driver when they started switching to uxa). I suppose > it's possible the horse power is going to the screen trying to keep up > with the scrolling text. I was in run level 3 and was using jhalfs, so the screen/netwrok traffic was minimal. I read somthing about the img file expanding as needed and that could slow things down. binutils was 332 seconds virtualized and 100 seconds on bare metal, so if that had held up, it would have been a 3x slowdown, not a 1.5x slowdown. -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
