Guys,
I was just at my accountant's home this morning. This is a freebie in my capacity of a professional computer consultant since 1988 and professional reviewer of AV software for Windows Magazine, then PC Magazine, and now CPU Magazine, since 1996.
Their kid's new PC is infected after only 45 days out of the box. It has the latest version on AVG free (which they installed, since they remembered I used to recommend it) and AOL's McAfee-based freebie AV tool. I didn't even know they bought this new PC, but at least they tried to protect it.
AVG isn't even complaining about the system being totally owned - it thinks everything just just peachy. McAfee is at least complaining about locating infected files that are starting every 90 seconds, but it can't clean them. I'll have a million laughs cleaning this box off over some evenings this week, but it's what I do. Professionally. For a living.
This is very typical of AVG free, which is why I've not been recommending it for the past 6 months, especially if the user doesn't have firefox which this machine doesn't.
You say you don't recommend any AV software for a gaming rig. To me, this nonsensical statement only proves to me that you really aren't testing these things very closely, are you?
One of the things I do when I review AV software, in addition to letting it try to deal with my virus zoo with more than 8500 active viruses, (with about 100 unique variants being added every month - the joy of running my own mailserver), is to run benchmarks, which get published with every review.
My standard test machine finishes booting XP in 29 seconds, and runs a Disk/CPU benchmark in 8:30 seconds, repeatably. Going from the notes I have in just this laptop, here are the following measurements for a few AV programs:
Norton AV 2008: boot: 49, Benchmark: 19:30
Norton Internet Security 2008: boot: 56, Benchmark: 22:40
CyberDefender Complete: Boot: 2:41, benchmark: DNF (stopped after 4 hours)
Eset NOD32: boot 31, Benchmark: 10:21
AVG Free 7.5: boot: 44, Benchmark 14:50
Kasperski 7.0: boot: 40, Benchmark 20:10
Eset NOD32 increases boot time from 29 to 31 seconds and increases the benchmark time by about 20%, sometimes less than half the load of other programs, all while doing very well my personal zoo (these stats are on my home machine - sorry) , as well as those from AV Comparatives foreground scan results (
http://www.av-comparatives.org/seiten/ergebnisse_2007_08.php) and zero-day heuristic results (
http://www.av-comparatives.org/seiten/ergebnisse_2007_11.php).
The AH framerate drop when running Eset Nod32, which is also on of my tests but it goes unpublished, is NEGLIGIBLE, being within statistical error (especially since there's no good, totally repeatable framerate benchmark in AH like there is for HL2 or UT for example). It isn't negligible for some other products, but I don't have those notes with this machine either.
I'm not trying to get into a pissing match either, but let's at least have some measurements and facts to back up our opinions - other readers are counting on this information being good.
Should anyone be interested, I'd be happy to report back here with what bugs I found on the kid's machine, and what versions of what programs let them slip through.
-Llama