Best way to compare xp builds?

N

Noncommittal

n00b
Joined
Jan 27, 2008
Messages
40
Right now I am working on optimizing my company's xp install image for better performance. We use Thinkpads and they come with all sorts of bloatware and crap they just don't need. I've been building a tweaked xp image to compare against our current one. I need some way to quantify the difference in performance. I ran the Sandra performance test on them and the final score was 2630 (my build) vs 2575 (old build). Unfortunately, I am not that familiar with Sandra and how much that difference in score really means. Any suggestions?
 
I’m not a benchmarking expert, but I don’t think than a synthetic benchmark like Sandra is really what you want. Synthetic benchmarks don’t generally do a good job in showing improved user experience that comes with cleaning up a machine. You might be better served by simply timing common activities of your users and showing before and after times.

My 2 cents. Good luck!
 
Yeah, Synthetic benchmarks aren't going to cut it.

You'll need to use practical tasks to determine the difference:
1. Login /logoff times
2. File copies of small, large, medium builds
3. Open Applications and use them. Bad bloated builds are easy to spot here.
4. Memory Utilization

It's hard to quantify a bloated build, but if you're using that crappy T series build from IBM, I promise you it's bloated. :) Startup and Shutdown should be a pretty good test for you.

I'm forced to use a bloated T series build. When I start my machine up, it's literally a 4 minute boot time another good 2 minutes to log in. On a good solid build, an XP machine should be into the domain and ready to go in ~1-2 minutes, depending on your logon script.
 
It isn't really performance that you are after to show a difference. I'd focus more on free memory and resources, reduced hard drive space usage, and the boot/load times mentioned already. That's where the crapware removal will pay off.
 
DeaconFrost said:
It isn't really performance that you are after to show a difference. I'd focus more on free memory and resources, reduced hard drive space usage, and the boot/load times mentioned already. That's where the crapware removal will pay off.
Click to expand...

Yes and to add to this, as an admin, you should know what they do and dont need to get their work done. Browse the add/remove programs and decide, "Do they really need this on their pc?". That should help alot. And since this is a workstation that will be imaged, all of the oter stuff that IBM puts into then thats considered "irrelevant", you might want to consider tossing that also.
 
Thanks for the input. Mostly I have been doing what was mentioned here. I've comparesd start up and shut down times, application open and close time, the number of running applications, memory usage, but I haven't tried any file copy comparisons.

I was hoping there was a synthetic that would provide me with a quantifiable difference, but I figured since they are the exact same hardware their would no significant difference. I will try bootvis and a few other things. Thanks.
 
Don't forget to run something like RegScrupXP or CCleaner when you've finished removing all the crapware. Majorgeeks has some software called PC-Decrapifier that should help in the removal as well.
 
You must log in or register to reply here.
Back
Top