“There’s no such thing as a free lunch.” If Windows users want to be secure, they need to pay for this privilege—money in some cases, performance in all of them. Security programs have been known to slow down any Windows PC—leaving users to struggle with a noticeably longer boot-up process, a less responsive workflow, slower gameplay, and reduced media application performance, among other things.

This is why, in recent years, antivirus companies have made an effort to communicate their products’ low footprints and minimal impact on PC speed. Symantec and Microsoft, for example, have included these aspects in their tag lines and on their product pages.

Performance is an important USP for users who want to be secure but don’t want their systems to be sluggish. The computing industry has also made many advances when it comes to raw processor power, storage speeds (think SSDs), and the amount of RAM shipped with new PCs. That’s why the TuneUp Blog team is constantly evaluating the situation—are security solutions still undoing the performance advances made by hardware manufacturers? Are they still bloated, running many services and processes, and thus making you work unproductively? In this series, we’ll compare current popular antivirus solutions and security suites in terms of their impact on PC performance. We’ll start off with this introduction, in which we explain the reasons for and methods behind conducting this comparison. In subsequent parts of this article, we’ll post the results of the AV products.

Article index

• Part 1 – Introduction (this blogpost)
• Part 2 – Microsoft Security Essentials results
• Part 3 – Norton Internet Security 2011
• Part 4 – Kaspersky Internet Security 2011
• Part 5 – AVG Internet Security 2011
• Part 6 – McAfee Internet Security 2011
• Performance Shoot-Out
• Part 7 – ESET Smart Security 4
• Part 8 – Avast Internet Security 6
• Part 9 – F-Secure Internet Security 2011
• Part 10 – BitDefender Internet Security 2012
• more parts to follow

At the beginning of this series, I would like to make one thing very clear: This test is designed to show how far security products have come these days and – perhaps – help make things better. However, being secure should be your number-one priority. Never make a decision about a security product based on just performance. There’s no doubt that performance is part of the decision-making process that goes into selecting the right antivirus scanner, but it shouldn’t be your primary criterion. Again, security comes first.

Why are antivirus solutions slowing down your PC?

One of the basic functionalities that come with current antivirus programs is real-time protection. It monitors when files and programs are being opened on your computer and alerts you to any suspicious activity. It’s important to know that these file operations happen all of the time: when you boot up, when you launch a program, when you download a file, etc. Most also feature an additional scan for e-mail attachments and Internet browser downloads.

Real-time protection is critical but also tends to slow down your machine. The monitoring process results in significant hard disk activity and processor load whenever your PC needs to perform hard disk operations. For example, my first test subject, Microsoft Security Essentials, produced more than 2,000 operations when I launched DVD playback software. Using Process Monitor, I tracked the activity of MsMpEng.exe, which is the process behind Microsoft Security Essentials.

This delays all operations—from starting a program to opening a file to switching between parts of programs. On newer machines, this might only be barely noticeable, or not noticeable at all. On older machines or on netbooks, however, this leads to quite a performance problem.

Test criteria and benchmarks

For testing purposes, I used a 2 GHz Windows 7 laptop with 2 GBytes of main memory. I regularly take this system on the road with me and use it for browsing the Web, e-mailing, and editing photos and basic videos.

For the benchmarks themselves, I took somewhat different approach as compared to our previous tests. I used a self-developed methodology called “scenario-based benchmarking”, which is a more real-world approach to testing PC performance. Instead of measuring single statistics, such as boot-up times, or applying benchmark programs to individual tasks, I performed a series of pre-defined actions—individually in succession—and measured how long it took to achieve the steps. Below I have outlined the steps I performed at very specific times to measure every single action.

1. 00:00:00 –Start/boot performance
   • Time it takes for the Windows 7 Desktop to appear
   • Time it takes for Windows Live Messenger to launch
   • Time it takes for Skype to start up
   • Exact boot-up time according to XPerf, a professional performance measurement utility that is part of the Microsoft Performance Toolkit
2. 00:10:00 – Resource check
   Ten minutes after boot up, I launched Windows Task Manager to detail the current memory load and the number of active processes.
3. 00:11:30 – Browser start
   One minute and 30 seconds after I checked Windows Task Manager, I started Google Chrome with five tabs and checked how long it took for the browser to fully appear.
4. 00:12:30 – Outlook launch
   Having closed Google Chrome, I waited two minutes and then measured how long it took to start Outlook, from clicking the application short-cut until the first message in my inbox showed up on the screen.
5. 00:14:30 – Opening PhotoImpact
   Next, I double-clicked on a huge 8 MByte picture I took on my last vacation and waited until it was fully visible in PhotoImpact. After it appeared on the screen, I waited 20 seconds and applied the “Enhance” filter to optimize the color and sharpness of the picture.
6. 00:16:30 – Opening PowerDirector
   I created a small video of my last vacation with Cyberlink PowerDirector, and used it to measure the exact time it took from double-clicking the PowerDirector project file to seeing the video-editing timeline and the preview of my clip. I also converted a small HD video clip into standard definition in AVI format.
7. 00:18:30 – Convert MP3
   I converted a two-hour audiobook (a 220 MB MP3 file) into AAC format using iTunes and measured the time it took for the conversion.
8. 00:20:30 – Copying large files
   Using my Wi-Fi connection, I copied a 184 MByte ZIP file from one PC to another. Since antivirus engines check all incoming and outgoing file transfers, I wanted to see how much of an impact they have on network traffic.
9. 00:25:00 – Extracting largefiles
   I extracted the ZIP file that was copied onto the hard disk. It contained about 300 files (a total of 550 MBytes), and required a lot of parallel work between the processor and hard disk drive, as well as the antivirus programs’ real-time protection.
10. 00:28:00 – Cinebench Results
    I launched Cinebench 11.5 and ran both the CPU-only and the OpenGL tests to see if and how well the computer handled real-time rendering with and without a full-blown antivirus scanner working in the background.
11. 00:37:00 – PCMark results
    Finally, I ran PCMark Vantage and ran the “Productivity” and the “Gaming” benchmark suites.
12. 01:00:00 – End

All of the steps and benchmarks performed touch on different areas of PC performance. Some were raw processor tests, while others relied heavily on Windows’ multitasking capabilities. I also tested file operations and used two synthetic benchmarks.

So why did I do all this? By running all of the tests at specific times one after another, I got a more realistic result. Users don’t start their computers, perform single tasks and then immediately shut down their systems. They run many programs simultaneously, copy files, close programs, and/or play games—all in a single session. The scenario-based approach shows a realistic load on Windows’ memory management.

However, to be on the safe side, I performed the scenario six times—three times without and three times with an antivirus scanner installed. Having performed the first three tests without an antivirus solution, I restored the PC to its original state using an image, so I was able to compare the results.

But enough specifics, let’s see how much of an impact we got with and without these programs.

Jump to the next part(s) in this series:

• Part 2 –Microsoft Security Essentials
• Part 3 – Norton Internet Security 2011
• Part 4 – Kaspersky Internet Security 2011
• Part 5 – AVG Internet Security 2011
• Part 6 – McAfee Internet Security 2011
• Performance Shoot-Out
• Part 7 – ESET Smart Security 4
• Part 8 – Avast Internet Security 6
• Part 9 – F-Secure Internet Security 2011
• Part 10 – BitDefender Internet Security 2012

What’s your experience with the current antivirus solutions out there? Are you seeing a performance loss on your PC? Do you want us to perform a test of a specific security software suite that’s out there? Let us know!

Do Security Solutions Slow Down Your PC? (Part 10 – BitDefender Internet Security 2012)