Vista – Benchmarking The Tweaks

I often get asked some version of the following question, “What benefit is there to tweaking Windows”. To Geeks the answer to this question falls along the lines of “We hold these truths to be self-evident”. In other words we know why and have been doing it so long that the answer is obvious to us.

Why We Tweak Windows

1: To make the computer faster. It never ceases to amaze me when some self-appointed expert says tweaking does nothing for your computer.

2: To customize the interface (tweaking is about more than just speeding up the computer itself). This can be for looks or for organization. Either of these can speed up your computing experience by allowing you to get things done faster. Not to mention making the computing experience more enjoyable.

The purpose of this article is simply to show that tweaking can indeed speed up Windows. Specifically, Windows Vista.

Why I Don’t Like Benchmarks

1 – You can never post a benchmarks without someone pissin’ and moanin’ about it. Accusations of manipulation, cheating, incompetence, etc.

2 – To properly benchmark a computer, especially an OS, is an extremely time consuming, tedious, and often boring process.

3 – I do not believe there is a 100% accurate way to benchmarks an entire OS. (see next comment)

4 – There is no benchmark to accurately reflect how you use a computer, app, game, etc.

So why benchmark? To try and measure the effect of the tweaks as best we can. Despite the limitations of benchmarks some folks need some way to measure beyond look and feel.

The Benchmarks

The Computer
Motherboard	Asus P4C800-E Deluxe
CPU	3.4GHz Extreme Edition
Memory	Corsair TWINX1024-3200C2PT
Video Card	Nvidia 6600GS
Sound Card	Audigy 2 ZS Plat
NIC	Onboard Intel Pro 1000
Hard Drives	Raptor 150 WD 120 Maxtor 300

Why this system? It is my utility machine and the one I could afford to be without for the several days of this test.

The Tests:

Boot Time

PCMark Vantage

Passmark

Methodology:

The entire process took several days ( I have a job and a life). Reasons for this methodolgy and tips to benchmark Vista can be found here: Measuring Performance in Windows Vista

1 – Wipe drive and Install Vista

2 – Install SP1 and all updates.

3 – Install benchmark tools.

4 – Run benchmark tools (to put them in to prefetch).

5 – Cleanup and defrag.

6 – Make an image.

7 – Restore from image, run process idle tasks, run defrag (so that both before and after tweaks will have the same disk layout).

8 – Run benchmarks and document.

9 – Repeat step 7.

10 – Tweak it. I did a complete walk-through of my Vista Tweak Guide just as it is written except that PCMark Vantage requires the visual settings the be for “best appearance” and the start menu must be default not classic. The Services settings were the Networked + Sharing settings except that the Diagnostic Policy Service was left at Automatic to measure boot times.

11 – Run benchmarks and document.

Boot Times
This can be found in the Event Viewer > Applications and Services > Microsoft > Windows > Diagnostics-Performance > Operational
Before: BootTime 46610
After: BootTime 38896

PCMark Vantage

The results are the best of three runs.

You can view these results online:

Before – http://service.futuremark.com/compare?pcmv=85098

After – http://service.futuremark.com/compare?pcmv=85739

Main test results	Before Tweaks	After Tweaks
PCMarkSuite	2370 PCMarks	2623 PCMarks
Memories Suite	1481	1564
TV and Movies Suite	1589	1603
Gaming Suite	1586	1611
Music Suite	2888	3183
Communications Suite	2357	2506
Productivity Suite	2015	2411
HDD Test Suite	4172	4708

Passmark

Best of three runs.

Test Name	Before Tweaks	After Tweaks
PassMark Rating	399.6	429.8

There it is. Tweaking your computer obviously has valid and measurable results. The margin of difference will vary by configuration and hardware.

Eric