We tested Mac OS X Leopard 10.5 and then later uniformly and gladly upgraded to 10.5.1 when that was released at the tail end of our testing cycle.

We installed Leopard Server on two different Xserve servers, one a dual PowerPC/G4-based server connected to an Apple X-SAN external 2TB RAID array, the other on the CoreDuo-based Xserve we tested in May. All machines were tested on a managed switch GBE IPv4/IPv6 network, and with an IEEE 802.11b/g/ Linksys WRT54g wireless access point.

We took measurements of initial operating system displacement based on Mac OS X Tiger 10.4.9 then benchmarked them using LMBench3. We then upgraded each server, noting the time that an in situ upgrade took. We then wiped the server of all software and noted the bare metal installation times.

Leopard Server's new applications were tested, including Time Machine, Wiki (and blog), and iCal functionality. Connectivity was tested to Microsoft Active Directory, NFS, OpenLDAP trees and Novell's eDirectory. We subjected the Leopard server to a variety of fingering tests using nmap, and also subjected the Apple firewall to assaults from a custom suite of typical firewall cracks and service fingering.

We also upgraded numerous Mac notebooks and desktops, using G4 CPUs (PowerBook G4 notebooks and PowerMacs) from 10.4.9 or 10.4.11 to 10.5 (then 10.5.1) in times ranging from 71 minutes to 184 minutes. Apple also supplied a MacBook Pro (2.4 Ghz Intel CoreDuo with 4GB of memory) for testing.

< Return to main test: Leopard improvements are spotty>