Re: [patch] voluntary-preempt-2.6.9-rc1-bk12-R6

[William Lee Irwin III]

On Iau, 2004-09-09 at 23:45, William Lee Irwin III wrote:
>> Something odd is going on, in part because I get *blistering* IO speeds
>> running benchmarks like dbench, tiobench, et al on tmpfs with striped
>> swap. In fact, IO speeds markedly faster than any other filesystem I've
>> ever tried, by about 30MB/s (i.e. wirespeed, where others fall about
>> 37.5% short of it). Virtual alignment issues do hurt, but the core
>> allocation algorithm appears to be better than good, it's astounding.

On Thu, Sep 09, 2004 at 11:11:39PM +0100, Alan Cox wrote:
> Thats a very atypical load where you can expect to get long linear write
> outs. The seek v write numbers for a disk nowdays have more in common
> with a tape drive. Paging tends to be much much more random.

Yes, I mentioned that those kinds of benchmarks are not the workload
we're shooting for in the second part of the message. The commentary
regarding dbench et al on tmpfs suggests that the lower-level parts of
the algorithm are sound, but are somehow driven inappropriately or in a
manner unaligned with what locality of reference there may be.


-- wli
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/