Greetings, all,
I thought you might be interested to hear about my experiences building Hackystat on my
new Mac. I have been interested in doing my Hackystat development on a Mac for quite a
while for a variety of reasons, but pre-Intel versions of the Mac (i.e. my G4 powerbook)
took roughly twice as long to build Hackystat as my WinTel machines, which simply wasn't
acceptable.
That does not appear to be the case any longer. I received a MacBookPro last week, and
got everything to the point where I could do some timings this morning.
My test was to do 'ant -q freshStart all.junit' on a Hackystat configuration consisting
of 25 modules (all 7 hackyCore_* modules, all 17 hackySdt_* modules, and the
hackyApp_StdCmd module).
I ran this command three times on three environments: the above MacBookPro (2.33 Ghz
Intel Core 2 Duo processor, 3GB RAM), my three year old IBM Thinkpad (X31, 1.6Ghz
processor, 1.25GB RAM), and my two year old desktop (Dell workstation P4 dual 3.4Ghz
processor, 2GB RAM). I'm using the standard Hackystat build environment (JDK 1.5.0_06,
etc.) on today's trunk.
The time in seconds to complete 'ant -q freshStart all.junit' for each of three
runs were:
Run1 Run2 Run3
Mac 186 186 183
Thinkpad 470 459 456
Dell 300 299 298
In contrast to my Mac, neither my Thinkpad nor my Dell are 'state of the art', but these
results still seem to indicate that the MBP is now a very acceptable choice for Hackystat
development.
If anyone wants to post their own timing data for this 25 module Hackystat configuration,
that would be great. I would be particularly interested in how long it takes on a more
recent WinTel or Linux box to see if the MBP continues to be competitive.
Cheers,
Philip
