What Remains of Magic Speed Improvements

Doesn’t every IT pro dream of finding a secret switch, a switch that increases clock speed, doubles the amount of RAM or makes the PC boot faster? Some do, and I admit I was fascinated when I heard that changing a simple boot option would significantly speed up startup time. I made some tests which confirmed what I had heard and blogged about my findings. A lot of people picked up that post, I was happy and everything was great - until I found out I was wrong.

When André from WinVistaSide commented on my article and insisted I was wrong I tested again. To my great shame I have to admit that he is right. Apparently this is a hoax (or put more mildly: a misunderstanding) even IT pros like myself fall for. Thank you, André, for pointing this out.

Standard Measures by Neil Cummings under CC

The Difficulty of Measuring

Measuring is easy, but measuring correctly is very hard. In German, physicists have a saying that illustrates this: “Wer misst, misst Mist.” [Who measures, measures rubbish.]

I had measured like this: first I timed the purportedly slow setting, booting twice, if I remember correctly. Then I changed to the “fast” setting and took another two or three measurements while booting. The latter numbers were 20% smaller. Proof the second setting is better? Obviously - if there are no additional external influences.

Instead of stopping here I should have reversed the setting a second time and taken more timings. I would have found that the numbers stay exactly the same. I know, because I did that just now. The “magic setting” has no measureable effect whatsoever.

But wait, there was a speed increase when I first enabled the “magic setting”. Yes, there was, but remember I mentioned external influences earlier? There are other powers at work that change the setup between timings, namely ReadyBoot, Windows’ boot optimization mechanism (here is a nice explanation).

Lessons Learned

Not measuring bullshit is difficult. Numbers lie. People, too, although in many cases unwittingly (as I did in this case). Numbers are used to simplify complex things into a single item. In many cases that is not possible without losing much of the detail and context, and in some cases the resulting number has no meaning at all.

Comments

Related Posts

The Impossibility of Measuring IOPS (Correctly)

The Impossibility of Measuring IOPS (Correctly)
If you have ever used Sysinternals’ Process Monitor, chances are high you were a little intimidated when you looked at your first capture: it probably contained hundreds of thousands of registry and file system events, generated in a minute or less. That amount of activity must surely indicate high system load - but strangely, very often it does not. Looking at the hard disk LED you will only see an occasional flickering, even though thousands of file system events are captured per second. How is that possible? Read on to find out.
Performance/Sizing

Latest Posts

Fast & Silent 5 Watt PC: Minimizing Idle Power Usage

Fast & Silent 5 Watt PC: Minimizing Idle Power Usage
This micro-series explains how to turn the Lenovo ThinkCentre M90t Gen 6 into a smart workstation that consumes only 5 Watts when idle but reaches top Cinebench scores while staying almost imperceptibly silent. In the first post, I showed how to silence the machine by replacing and adding to Lenovo’s CPU cooler. In this second post, I’m listing the exact configuration that achieves the lofty goal of combining minimal idle power consumption with top Cinebench scores.
Hardware

Fast & Silent 5 Watt PC: Lenovo ThinkCentre M90t Modding

Fast & Silent 5 Watt PC: Lenovo ThinkCentre M90t Modding
This micro-series explains how to turn the Lenovo ThinkCentre M90t Gen 6 into a smart workstation that consumes only 5 Watts when idle but reaches top Cinebench scores while staying almost imperceptibly silent. In this first post, I’m showing how to silence the machine by replacing and adding to Lenovo’s CPU cooler. In a second post, I’m listing the exact configuration that achieves the lofty goal of combining minimal idle power consumption with top Cinebench scores.
Hardware