August 7th, 2008

VIPRE performance shoot-out - Does it really not slow down your PC?

Posted by Adrian Kingsley-Hughes @ 8:19 am

Categories: On the test bench, Performance, Security, Software

Tags: 32-bit, 32-bit system, performance, pc, norton co., antivirus, symantec norton antivirus, ram, 64-bit, viruses and worms

[UPDATE: Many of you have asked about the effectivness of VIPRE. Testing  security software is a big job and well outside of my capability. However, VIPRE has passed West Coast Labs' Checkmark certification. Details here. I am told that more certification is in the pipeline.]

The other day when I blogged about Sunbelt Software’s latest VIPRE antivirus solution I took the claims that it doesn’t slow down your PC pretty much at face value. Time to see if there’s any validity to these claims.

Test system #1

• Quad-core Intel Core2 Extreme QX9650 3.0GHz 
• DDR2 800 RAM (2GB, 4GB and 8GB)
• WD Raptor 150GB hard drive
• Vista Ultimate 64-bit SP1, fully patched up

Test system #2

• Pentium 950D processor (3.4GHz)
• DDR2 667 RAM (1GB, 2GB and 4GB)
• 250GB Maxtor hard drive
• Vista Home Premium 32-bit SP1, fully patched up

Note: When loaded with 4GB or RAM the 32-bit system registered 3.2GB of RAM.

Note: No optimizations were carried out on either system other than defragging the drive and making sure that all background tasks had finished.

Testing

Each system was imaged so that three antivirus products could be installed onto the OS (after testing the system was rolled back to the clean image).

The three antivirus products tested were:

• VIPRE
• Kaspersky AntiVirus 2007
• Norton AntiVirus 2008

All three products were fully updated before testing.

Note: All settings on each of the antivirus software packages was left in the default configuration.

Testing was done with PassMark PerformanceTest 6.1 (32-bit and 64-bit for the respective operating systems). For each test PassMark PerformanceTest 6.1 was run three times and the scores averaged to achieve the final score.

Note: Only the effect of background (or realtime) scanning was tested. Running the benchmark during a full system scan generated results that were too inconsistent to publish.

Results and Conclusions –>

Pages: 1 2