diglloyd Mac Performance Guide
The Speed You Need
The Speed You Need
Photography Blog and Publications
Photography Blog and Publications

Mac Pro Nehalem Tests: Software Development

Last updated June 01, 2009 - Send Feedback

This page tests software developer tests — building applications.

XCode build

Mac Pro Nehalem memory copy speed bandwidth
CPU Utilization
16 cores @ 100%

Build diglloyd DiskTester using Apple XCode. This test scales to large number of CPU cores extremely well and is not sensitive to disk speed at all.

Here the MP08 takes 75% longer than the MP09, not surprising given that all 16 cores of the MP09 are nearly pegged at 100% usage (test duration is green band as seen at right).

This is a fantastic performance by XCode, and it shows what a well-written program can do with 16 cores.

Single vs dual CPU: XCode scales beautifully; any developer using XCode is ill-advised to use a single-CPU Mac Pro: the dual-CPU MP09 is twice as fast.

 

Mac Pro Nehalem memory copy speed bandwidth
Apple XCode — build DiskTester

Java/Maven build — GlassFish V3

Mac Pro Nehalem memory copy speed bandwidth
CPU Utilization
Very poor.

This test does not scale well at all to multiple cores because it is a single-threaded affair, and also a CPU-bound test, so top speed of a single core is paramount.

Gains here are decent, but not exciting— only 7% for our estimated 3.2GHz MP08.

Single vs dual CPU: The single-threaded maven build makes no use of extra cores, so build times are the same for the single or dual CPU MP09.

Mac Pro Nehalem memory copy speed bandwidth
Maven build Glassfish V3

Conclusions

Coming soon...

 

See How to Choose and Buy a Mac.

Up to 8TB of Thunderbolt Storage!
Up to 8TB of Thunderbolt Storage!
diglloyd Photography, Subscribe Now!
diglloyd Photography, Subscribe Now!

diglloyd.com | Terms of Use | PRIVACY POLICY
Contact | About Lloyd Chambers | Consulting | Photo Tours
Mailing Lists | RSS Feeds | Twitter
Copyright © 2008-2015 diglloyd Inc, all rights reserved.