By Fred Langa
A reader’s complaint about Microsoft Security Essentials (MSE) spurred head-to-head comparison tests of AV-software resource usage.
I put six popular, free antivirus tools through their paces and measured their impact on startup and shutdown times, disk space, and RAM use.
Normally, reader letters appear in the LangaList Plus section of this newsletter. But sometimes, one comes along that warrants treatment in greater detail.
One such letter is this one, from subscriber Bill Garfield:
- “Occasionally you either recommend or suggest Microsoft Security Essentials as a viable free alternative to commercial anti-malware programs without mentioning the also-free competing products such as AVG and Avast, to name just two.
“MSE is an enormous resource pig, adding a full 30 seconds or more to boot times. Many Windows users look to Windows Secrets for tips and tricks to improve performance. Based on my experience, MSE does more to hobble overall system performance on supported platforms than to improve it.
“Please stop recommending MSE, or at least include an advisory disclaimer cautioning users that loading MSE has been reported to cause performance to suffer.”
As the Windows Secrets author who has recommended MSE more than any other contributor, I was alarmed by Bill’s letter. I decided to run a series of head-to-head tests, comparing MSE’s resource usage to other popular, free antivirus tools.
I did not test these tools’ ability to detect malware. With the possible exception of ClamWin — which is quite new and still evolving — most third-party tests rank all these tools as acceptable and some of them as excellent. (Valid malware detection-and-removal testing can be done only in specialized labs, and their published results are difficult to compare.)
Because MSE is completely free, I chose to compare it against programs that also are truly free — no strings (such as free trials of commercial programs) attached.
I chose software that, like MSE, is specifically designed for malware protection — I didn’t include integrated, do-it-all security suites.
To make the selections, I used popularity ratings from several sources, such as the download stats and user ratings from CNET, MajorGeeks, and other sites. (The rankings and numbers in the list below were current at the time of writing.) I also gave extra weight to products frequently mentioned by readers and other authors in the Windows Secrets newsletter as well as tools that looked especially promising (as you’ll soon see).
I settled on these six antivirus tools:
- Microsoft Security Essentials (site): Because I’ve recommended Microsoft’s consumer anti-malware application in previous stories, I’ve made it the baseline for these tests.
- Avast Free Antivirus (site): Avast claims its package is the “world’s most popular antivirus.” CNET also ranks it #1 on its download list.
- Avira Free Antivirus (site): It ranked #2 among CNET users, and it scored an impressive 4.7 out of 5 rating on MajorGeeks.
- AVG Technologies’ AVG Anti-Virus Free (site): The publisher of this app says it’s used by “over 100 million people.” CNET ranks it at #6 in popularity.
- Comodo Antivirus (site): Comodo gets very high users rating on both MajorGeeks and CNET.
- ClamWin Free Antivirus (site): Not as well known as the other AV packages, ClamWin is unique; it’s a free, open-source software project released under the Free Software Foundation’s GNU General Public License.
I know there are many other products out there, but I couldn’t test them all. I believe the preceding list is a good sampling.
Designing resource usage tests for AV products
Setup: To produce this comparison, I used Oracle’s VirtualBox (site) to set up a fresh, clean, fully up-to-date Windows 7 SP1 installation in a virtual PC (VPC).
Next, for reasons I’ll explain in a moment, I added two pieces of software: Piriform’s CCleaner (site) and a custom batch file that invoked the advanced mode of Windows’ built-in disk-cleanup tool, cleanmgr.exe. (The advanced mode is more thorough than the standard cleanup. You can read about the batch-file method of using cleanmgr in the Nov. 10, 2011, Top Story, “Putting Registry-/system-cleanup apps to the test.”)
I then cloned (copied) that initial, virtual-PC setup five times, ending up with six identical virtual PCs.
On each virtual PC, I installed one of the six AV tools, accepting whatever default settings the apps set at installation. When prompted, I allowed the software to update itself and run an initial, post-installation scan.
Next, I rebooted each VPC to make sure the setup was 100 percent complete and running normally.
I then deleted the installation file(s) and cleaned the system of any other temporary files created during download and installation, using the cleanmgr batch file and CCleaner. This step made sure nothing left over from the setup would affect the tests.
Testing the AV tools: To compare any effects on system startup and shutdown time, I powered off and powered on each VPC three times, using a stopwatch to track how long each start and stop took. The times given below are the averages of the three runs.
To measure the amount of disk space each of these apps occupies, I used Windows Explorer to view the properties of the C: drive on each virtual PC. I noted the amount of disk space available before and after installing each anti-malware app.
For RAM use, I started Task Manager in each system, waited five minutes for the system to fully settle down, and then noted how much RAM was in use before and after the apps were installed.
The results of these tests appear in the following tables.
Measuring the change in startup times
Windows’ startup happens in two parts: the initial system bootup before the sign-in prompt, then the time Windows takes to load user settings (from sign-in to the full appearance of the desktop).
I started timing when I launched the VPC, then paused the stopwatch when the Windows logon prompt appeared. After entering my user name, I simultaneously hit Enter and restarted the stopwatch. I kept the stopwatch running until the notification area was fully populated, all subsystem icons (sound, networking, and so on) were up and active, and all desktop icons appeared. Table 1 shows the results.
Table 1. In this and the following tables, the category’s best
result is highlighted in green and the worst in red.
As you can see, the open-source ClamWin offered the fastest average startup time (about the same as starting up the PC without AV software), closely followed by MSE. Avira had a significant impact on startup — more than double the fastest three products.
In this test setup, MSE doesn’t have any real impact on startup time. In a real-life situation, very few PC users will notice the one-second difference between ClamWin’s 35-second boot and MSE’s 36-second boot.
On the other hand, Avira’s 83-second average boot is quite noticeable. In fact, Avira’s boot was so slow, I thought something was wrong with the setup and so did it over from scratch. But the results were consistent — consistently awful.
Measuring the effects on shutdown times
Shutdown timing was simple: I simultaneously clicked Shutdown and started the stopwatch. I then stopped the clock when the VPC session shut down. Table 2 shows these results.
Table 2. ClamWin has the fastest shutdown time.
Although there were differences in shutdown times, they were much smaller than with the startup times — too small to worry about. ClamWin again was the fastest; its eight-second time stood out among the six apps. At 14 seconds, Comodo was the slowest — but it was only three seconds slower than MSE and Avast, the two second-place finishers.
Avira records a sizable RAM footprint
To make the RAM-utilization numbers easy to digest, I used MSE’s results as the baseline. Table 3 shows how much RAM — more or less — each of the other five tools used, in megabytes.
Table 3. RAM use varied significantly.
Avast consumed the least amount of RAM — 13MB less than MSE. AVG and ClamWin were on par with MSE, but Avira used a whopping 139MB more.
Disk-space footprint (disk space used)
As with RAM use, I set MSE’s disk footprint as the baseline. Table 4 shows the other programs’ disk use in gigabytes relative to the MSE baseline.
Table 4. Disk-space use varied only negligibly.
Unless your hard drive is near capacity (in which case you have more pressing problems than the AV software footprint), there are really no significant differences among the six products. Comodo used just 0.6GB less disk space than MSE, and AVG took up only 0.2GB more. In today’s era of 500GB and larger drives, disk space use should not be a factor in picking one of these AV products over another.
Summing up antivirus-software resource use
The immediate conclusion — at least in these controlled-environment tests — is that MSE is not the “resource pig” some PC users think it is. In fact, it offers respectable, near-best numbers in every category. Whatever was going on with Bill’s system is not likely to be intrinsic to MSE.
Table 5. Here are all the results, for easy side-by-side comparison. The best results are shown in green, the worst in red.
If there’s one app that consumes more PC resources than its competitors, it’s Avira — with the heaviest RAM use and significantly slower startup time.
ClamWin is a pleasant surprise; it performed well in every category and earned two “best of breeds.” However, I’m reluctant to recommend it because it’s a relatively new product, is used by a relatively small number of people, and is still in Version 0.9x release as of this writing.
The bottom line: My conclusions — and yours
For me — for now — nothing in these numbers would alter my recommendation — and personal use — of Microsoft Security Essentials. It’s free, it’s in widespread use, and it has proven itself in the real world. And on most systems, it has little effect on system resources.
That said, no software is ideally suited for every configuration. If one anti-malware tool doesn’t work well on your specific setup, then it makes perfect sense to try some other AV app. You have lots of options.
I hope the performance numbers I’ve reported may help steer you to the tool that offers what you seek — faster boot and shutdown, lower RAM use, and smaller footprint.
| Feedback welcome: Have a question or comment about this story? Post your thoughts, praise, or constructive criticisms in the WS Columns forum.|