> Is this a known problem with 2.4.17 and/or the 3.5 GB userspace patch? I
> have not tried turning off the 3.5 GB config option (`CONFIG_05GB'). I do
> have `CONFIG_MK7' set.
The configuration I would try first on 2.4.19pre1aa1 with 2.5 GB of RAM is
CONFIG_3GB=y and CONFIG_NOHIGHMEM=y. If that causes some other problem,
I'd go with CONFIG_2GB, then finally CONFIG_1GB. Each config changes
the user/kernel memory split. The loads I've run suggest for best
performance:
CONFIG_2GB > CONFIG_1GB > CONFIG_05GB
On my 1GB box, I've been running CONFIG_2GB for a couple months and
performance is great. Only lingering question for me is the handling
of oom. Oom hasn't caused be a problem, but I haven't seen the oom
killer do it's work when I create an oom condition.
BTW, Andrew Morton's read_latency2 patch is a winner. The read_latency2
diff I use on my 2.4.19pre1aa1 boxes is at:
http://home.earthlink.net/~rwhron/kernel/read_latency2-2.4.19pre1aa1.diff
-- Randy Hron- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/