Plone on the MacBook Pro

[Andreas Jung]

--On 3. März 2006 13:40:13 -0800 Alexander Limi <limi@...> wrote:

> Hi,
>
> Since the Plone developers have an (unhealthy? ;) obsession with Apple
> laptops, I got the Apple people to install Xcode on a MacBook Pro in the
> store today, and performed some basic benchmarks.
>
> If you want the short version and the pretty (for the MacBook, at least
> ;)  graph:
> http://limi.net/media/macbook-plone.jpg
>

Nice, the numbers are impressing. The poor startup performance is likely
caused by the poor performance of the HFS for seek operations. The ongoing
eggification process might reduce the startup time as well. For a project
we have put everything below lib/python inside a .ZIP archive and Zope on
Windows runs very fast with this approach (although it required some
changes in same packages).

-aj

[Alexander Limi]

Hi,

Since the Plone developers have an (unhealthy? ;) obsession with Apple  
laptops, I got the Apple people to install Xcode on a MacBook Pro in the  
store today, and performed some basic benchmarks.

If you want the short version and the pretty (for the MacBook, at least ;)  
graph:
http://limi.net/media/macbook-plone.jpg

Test setup
==========

Hardware
--------
MacBook: 2GHz Core Duo, 1GB RAM, 5400rpm disk
PowerBook: 1.5GHz G4, 7200rpm disk
(both disks are Seagate Momentus, same generation)

Software
--------
Latest Xcode (with gcc4)
Zope 2.8.6
Plone 2.1.2


Zope 2.8.6 compile
==================
MB: 49s
PB: 2m 3s

First startup
=============
MB: 48s
PB: 56s

(after dropping in Plone, includes language file compilation and other  
first-run tasks)
Startup is measured for 'bin/zopectl fg', which runs Plone in debug mode.


Subsequent startups
===================
MB: 11s (!)
PB: 37s

Unit test run (1567 tests)
==========================
MB: 152s
PB: 355s


Tests were run using the "time" command, and rounded to nearest whole  
second.

I didn't run the tests a lot of times, but I did of them twice, and the  
differences weren't statistically significant.


Other interesting observations
==============================

- In all of these cases, one of the cores was mostly idle, and you could  
for example run the unit tests for Zope 2.8 and Zope 2.9 in parallel in  
the same period of time with almost the same performance (did not get to  
test this, though - there will be competition for I/O and memory access,  
obviously).

- The MacBook Pro didn't ever turn on its fan while doing these tests -  
the PowerBook did.

- The PowerBook in this test has a 7200rpm disk, and the MacBook a 5400rpm  
disk. When I exchanged my 4200rpm disk on the PowerBook with a 7200rpm  
one, Plone startup went from over a minute to 37 seconds. I assume this  
would help the MacBook too, since starting Plone accesses a lot of files.

- Hanno's AMD 64 Dual Core 3800+ desktop machine only runs the unit tests  
5 seconds faster than the MacBook Pro (about 1% perfomance difference ;)

--
_____________________________________________________________________

      Alexander Limi · Chief Architect · Plone Solutions · Norway

  Consulting · Training · Development · http://www.plonesolutions.com
_____________________________________________________________________

       Plone Co-Founder · http://plone.org · Connecting Content
   Plone Foundation · http://plone.org/foundation · Protecting Plone



-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642
_______________________________________________
Plone-developers mailing list
Plone-developers@...
https://lists.sourceforge.net/lists/listinfo/plone-developers