Messages in this thread | | | Date | Thu, 28 Jan 1999 20:12:48 +0100 | From | "Dr. Werner Fink" <> | Subject | Re: MM deadlock [was: Re: arca-vm-8...] |
| |
> > Ok its a bit better than a single PII 400 MHz :-) > > ... with less than 64MB the break downs are going to be the common state > > whereas with 128MB the system is usable. Nevertheless whenever both make > > loops taking the filesystem tree at the same time, the system performance > > slows down dramatically (a `break down'). > > Not for me. That's probably just the advantage of having swap on a > separate disk, but I've got both a "find /" and a "wc /usr/bin/*" > running right now, and interactive performance is not noticeably > degraded on 2.2.0-release with 64MB (and that is with two active users > on the box right now). Concurrent filesystem and swap IO on the same > spindle is always going to suck.
I'm not talking about a simple find, ... the two "make MAKE='make -j10'" in /usr/src/linux/ and /usr/src/newkernel/linux/ do force this `break down' with 2.2.0-pre9 if the two makes are entering /usr/src/linux/fs/ or /usr/src/newkernel/linux/fs/ respectively at the same time which increases the load a `bit'.
Werner
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |