Performance Shoot-Out: Do Security Solutions Slow Down Your PC?

It’s an age-old PC fact: Protection from malware, intruders, hackers, phishing and even spam comes with a loss in performance. We asked ourselves—is this still true in today’s world of super-fast computers? Have security suite makers learned any lessons from the past and reduced the footprint of their products? Over the past several months, we conducted extensive performance tests (and will continue to do so on a regular basis) to find out.

We picked the most popular security products as well as listened to your suggestions. So far, we’ve tested the following solutions:

• Part 1 – Introduction/Test criteria
• Part 2 – Microsoft Security Essentials (MSE) 2.0
• Part 3 – Norton Internet Security 2011
• Part 4 – Kaspersky Internet Security 2011
• Part 5 – AVG Internet Security 2011
• Part 6 – McAfee Internet Security 2011
• Part 7 – ESET SmartSecurity 4

Important disclaimer

Again, we’re sharing our findings because we want to help you make good choice regarding your security product. As we previously stated, our focus is on testing the effectiveness of these solutions. Don’t select product A (with a low virus detection rate) just because it doesn’t affect your PC’s performance as much as product B (with a higher virus detection rate). However, if you’re deciding between two products with a 99.8% detection rate, you should, of course, go with the one that’s faster!

How we test

To accurately measure performance, we compared the speed of 20 operations (divided into ten benchmark categories, such as “Boot performance” or “Open up a PowerDirector project”) that show the impact of each security solution on your PC. We’ve used a MacBook Air (2009, Core 2 Duo with 2 GB of RAM) for performing these tests. Ready? Let’s dive in!

Timecode 00:00:00 – Scenario starts: Boot performance

We measured boot performance in four stages. First, we calculated the time it took for the Windows desktop to appear, Skype to launch and Live Messenger to fire up. Then, we looked at the total boot time as reported by the Microsoft Performance Toolkit.

We used XPerfMgr to determine what kind of an impact the security suites had on our test machine.

Overall, ESET, McAfee Internet Security and MSE slowed things down the least, followed by Norton Internet Security. ESET won this round, as it had next to no effect on boot time.

Timecode 00:10:00 – Resource check

Ten minutes after boot up, we launched Windows Task Manager to detail the current memory load and the number of active processes. Looking at the results, only AVG and McAfee added a significant amount of processes and memory usage. ESET again won this round. This benchmark shows that the developers of the other solutions have done their homework and tried to reduce their footprints.

An average of 70-80 MB of memory usage seems like a lot at first, but in the days of 2, 4 and even 8 GB of RAM, it’s barely noticeable. Besides that, Windows Task Manager revealed that Microsoft Outlook and Internet Explorer 9 (with seven tabs running) eat up almost three times more memory than the currently installed Microsoft Security Essentials 2.0.

In terms of raw memory usage, security suites are the least of PC users’ problems. These results would have looked a lot different five years ago.

Timecode 00:11:30 – Browser start-up

We used Google Chrome to determine if security suites had any impact on start-up times.

The results? Google Chrome always launches in exactly two seconds. However, the moment we installed security software (no matter the solution), the web browser needed an additional second to load. Fortunately, that is barely noticeable, and the delay only occurred during the first start; with all subsequent loads resulting in significantly faster times.

Timecode 00:12:30 – Launching Outlook

Next, we measured how long it took to start Outlook from clicking the application shortcut until the first message in our inbox showed up on the screen.

Oddly enough, MSE was the only program that didn’t delay Outlook at all—perhaps Microsoft’s products favor one another?! Other security suites noticeably affected this application’s start-up performance.

Timecode 00:14:30 – Opening an 8 MB photo in PhotoImpact

We then double-clicked on a huge 8 MB picture and waited until it was fully visible in PhotoImpact. After it appeared, we waited 20 seconds and applied the “Enhance” filter to optimize the color and sharpness of the picture. So, how long did that take?

Performance in PhotoImpact suffered just a bit. While all of the contenders added a few extra seconds to both loading and editing the photo, AVG interfered a bit more than the rest taking almost twice as long to apply the “Enhance” filter.

Timecode 00:16:30 – Opening up a PowerDirector project

Next up was our video-editing performance test. For that, we created a small vacation video with Cyberlink PowerDirector, and measured the exact time it took from double-clicking the project file to seeing the video-editing timeline and preview.

We also converted a small, high-definition (HD) video clip into standard definition (SD) in AVI format. So, what kind of an impact can you expect from your security suite?

It is amazing how much of an impact the security suites had on the start-up of PowerDirector. In some cases, it needed three times longer. However, MSE was the only solution that affected the HD to SD conversion. ESET was the only solution to not cause any slowdown.

Timecode 00:19:30 – Converting a large (220 MB) MP3 file into AAC with iTunes

What better way to measure raw CPU power then by converting a two-hour audiobook (a 220 MB MP3 file) into AAC format using iTunes? How did our test candidates fare?

Ouch! Kaspersky and AVG added 30-45 seconds to the file conversion. This is quite noticeable—we expect security solutions to stay in the background while we need to perform resource-intensive tasks. McAfee was the only one that didn’t impact the conversion. Bravo!

Timecode 00:30:30 – Copying a large file over a network

Using a Wi-Fi connection, we copied a 184 MB ZIP file from one PC to another. Since anti-virus engines check all incoming and outgoing file transfers, we wanted to see how much of an impact they have on network traffic.

Surprise! McAfee’s and ESET’s solutions were the only programs we tested that didn’t affect network file transfer performance. All of the other tools took another 10 seconds to complete this action.

However, we should note that the version of McAfee we tested did not include a network scanner, while others did.

Timecode 00:35:00 – Extracting a large file

Parallel computing performance was also put to the test. We extracted the ZIP file that was copied onto the hard disk. It contained about 300 files (a total of 550 MB), and required a lot of parallel work between the processor and hard disk drive, as well as the security suites’ real-time protection.

The results? Security suites need to protect users from opening up malicious files that often hide inside compressed folders, such as ZIP or RAR. In most cases, the extraction took between 5-10 seconds longer. One exception to this was McAfee Internet Security 2011. We did not there was no noticeable effect on the RAR file extraction with this solution.

Timecode 00:38:00 – Cinebench performance

The next question we had: Do security solutions interfere with rendering performance? To answer that, we 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 security engine working in the background.

(More points equals better performance)

Here’s what we found— Only MSE and Kaspersky dropped Cinebench’s performance a bit. The other security solutions let the benchmark software do its 3D rendering of your software mode (CPU) and the graphics card (OpenGL).

Timecode 00:47:00 –PCMark results

Finally, we ran PCMark Vantage and used the “Productivity” and the “Gaming” benchmark suites.

(More points equals better performance)

All of the products we tested had an impact on PCMark’s performance. As we previously mentioned, McAfee produced different results each time we ran the tests and even crashed PCMark once. This is why we excluded McAfee from this section.

The key PC performance results

There is no question about it—security solutions slow down your PC. Fortunately, most modern security suites will not actually “kill” performance like their predecessors did. Out of the five products we tested, AVG had some effect on PC performance. Microsoft and Kaspersky followed with a noticeable impact on boot and application performance. And ESET, Norton and McAfee have so far done the best job. However, our series is far from over! We will continue to perform tests and update this blog post each month.