[sclug] Where to start troubleshooting machine speed?

[Alex Butcher]

On Mon, 5 Sep 2005, Darren Davison wrote:

> I've got a desktop box running Gentoo that I've had for a while and also
> recently installed Gentoo on a laptop that I use for work.  For some while
> I've been suspicious of the general speed of the desktop but put it down
> mainly to the graphics chipset (Intel 855G) not being much cop.
>
> The laptop however uses the same graphics chipset and has the same drivers
> installed as the desktop,

They're the same drivers, but are they the same versions? As you're using
Gentoo, were all your packages compiled in the same order, using the same
version of the compiler and with the same flags and features enabled (or,
preferably, exactly the same binaries)?

If the answer to any of those questions is 'no' or 'don't know', you're not
performing an apples-to-apples comparison. Getting the same binaries on both
machines should be the first step in any investigations you make; probably
the most convenient way for you to do this would be to use Knoppix (or some
other bootable live CD) temporarily, but installing a binary distro (e.g.
Fedora, CentOS, SuSE) with the same errata packages installed would also be
acceptable.

> but X apps noticeably all run much faster on the laptop.  It has more
> memory than the desktop (2G vs 768Mb) but a slower processor (1.6Ghz vs
> 2.4Ghz).  Funnily enough though, it also seems significantly quicker
> compiling software than the desktop box.

Things I would look at:

- Disc DMA, interrupt unmasking, 32-bit IO (significant for compilation,
amongst other things, I find)

- Check physical configuration of desktop machine (e.g. cabling, jumpers)
especially if a DIY box (though I've had problems with largeish PC vendors
misconfiguring the machines they've sold)

- X configuration (for X application performance)

- Booting the laptop's kernel with mem=768M to make its configuration more
closely match that of the desktop.

- Comparing CPUs; cache and frontside-bus speed for a start.

- Running CPU benchmarks to verify that the 2.4GHz machine has a
better-performing CPU than the 1.6GHz machine (remember that a CPU with a
better architecture running at a lower clock speed can outperform poorer
architectures running at faster clock speeds!)

- Check heatsink/fan in desktop; remember that Intel CPUs slow down if they
overheat as a first step in protecting themselves from damage.

> It's all subjective of course, I have no benchmarks to offer, but the
> difference in general speed of operation is very, very noticeable.
>
> The desktop always seems to have plenty of RAM free and it's certainly not
> doing any swapping.  Any hints on where I should look to start figuring this
> speed differential out?  Could it all really be just due to the extra RAM in
> the laptop?
>
> Regards,

HTH,
Alex.
-- 
Alex Butcher      Brainbench MVP for Internet Security: www.brainbench.com
Bristol, UK                      Need reliable and secure network systems?
PGP/GnuPG ID:0x271fd950                         <http://www.assursys.com/>