As a matter of fact, back in the time when I was using a single-core Intel P4 CPU, I also tried out different partition formats, like ReiserFS, XFS, JFS... and never got a significant enhancement regarding the here-debated problem (though they are real nice file systems) (just they lack the wonderful capability of ext3 to recover the files from almost any sort of crash). So there would be something real special about ext4.
I read some online texts about ext4 but there was nothing obvious as to why ext4 simply allows the kernel to behave correctly. It is claimed that ext4 is also aimed at mission-critical applications, which leads for example to timestamps with a precision of a nanosecond, compared to a precision of a second for ext3. Using the second as base unit, sure will prevent the kernel from making a fluent scheduling. But I have no idea whether there is a link between the timestamps in the hard disk memory units and the kernel task scheduling... Maybe another of the mentioned new features of ext4 has the magical effect... Did somebody try to use an ext3 formatted partition but mounted by the ext4 software? ext4 can do that, but what will be the impact of this on the problem? I suppose mentioning the partition as ext4 instead of ext3 in /etc/fstab is enough but I didn't try this out. -- [jaunty] cpu scheduling is not optimized for multitask https://bugs.launchpad.net/bugs/363663 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
