Re: nifty swapspace trick

On Mon, 29 Oct 2007, Jamie Hushower wrote:

> A "typical" flash drive will not exceed the read/write speed of
> a "typical" SATA drive. The theoretical *maximum* of a flash drive on USB
> 2.0 is 60MB/s, but I suspect 30MB/s is more realistic for typical drives.
> A typical 7200rpm SATA drive might average 45MB/s. If you have relatively
> low disk I/O (though high RAM usage) to begin with, you might be losing
> speed by using a flash drive for swap space. I'm curious what Microsoft
> has to say on the matter, but am too lazy to look it up. Can anyone offer
> better numbers than I have or refute it?

Comparing storage devices on different buses is definitely a problem in
itself. So, in interest of intellectual honesty, how about just comparing
flash versus mechanical magnetic media on the same bus?

The reality is that both outperform each other under different workloads.
Recent flash memory types have worse write performance, on-par read
performance, and tremendously better seek performance. The latter is
especially important in the real world, especially when filesystems and
multiple process concurrent I/O can destroy contigous reads and writes on
mechanical systems.

I personally don't get the interest in SSD drives or flash memory outside of
very limited, specialized applications. A predominantly read-only database
makes a good candidate for them (much higher MTBF when you eliminate write
cycles), and consumers will like never having to reboot their laptops, just
do a fast restore and roll. But beyond that?

When MTBF isn't so horribly skewed by reasonable real-world read-write
ratios, I'll take it more seriously.

         --Arthur Corliss
           Live Free or Die
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Mon Oct 29 13:47:07 2007