System Performance Impact

Discussion in 'other anti-virus software' started by SourMilk, Mar 14, 2007.

Thread Status:
Not open for further replies.
  1. SourMilk

    SourMilk Registered Member

    Joined:
    Mar 31, 2006
    Posts:
    630
    Location:
    Hawaii
    With all the discussion about detection rates and cleaning abilities of certain antiviruses, I thought it would be interesting to see how some antiviruses effected system performance. Here are my surprising findings:

    Assumptions: No driver changes were made between tests; all tests were conducted after updating signatures and a reboot; all tests were started on a fresh (FD-ISR) install; no changes were made to the configuration of the antiviruses - they were installed as downloaded.

    Disclaimers: These are the results using my high-end computer so your results may vary. I only chose free editions of popular antiviruses - if I missed "brand X", I apologize; configuration of antiviruses for your machine may have an impact different than those presented.

    No Antivirus running = 100 pct on total system benchmark and hard drive benchmark

    Number 1 is Avast with 99.7 pct on total system benchmark and 99.0 pct on hard drive benchmark

    Number 2 is AVG AV with 99.4 pct on total system benchmark and 98.5 pct on hard drive benchmark

    Number 3 is AOL AVS with 97.8 pct on total system benchmark and 80.9 pct on hard drive benchmark

    Number 4 is Antivir with 97.8 pct on total system benchmark and 73.8 pct on hard drive benchmark

    I would suppose one would have to consider the detection rate/cleaning abilities of an antivirus along with the system performance impact when choosing one of the above.

    Personal note: I always thought AVG was lightest on my machine but Avast showed me different. I especially thank First Defense-Instant System Recovery and CrystalMark for their outstanding programs.

    SourMilk out
     
  2. C.S.J

    C.S.J Massive Poster

    Joined:
    Oct 16, 2006
    Posts:
    5,029
    Location:
    this forum is biased!
    Dr.Web 100 pct / 100 pct :D

    i aint too sure about avast being top-dog on the ones you tested, avg outclasses it in my experience for pc performance.
     
  3. WSFuser

    WSFuser Registered Member

    Joined:
    Oct 7, 2004
    Posts:
    10,632
    what settings did u use for each AV and why only free AV?
     
  4. Bluenile

    Bluenile Registered Member

    Joined:
    Feb 2, 2007
    Posts:
    122
    Location:
    UK
    I find AOL AVS runs lighter and faster than Avast on XP Home SP2 (It's currently using 5.6Mb of RAM for its 2 processes as I type). It has better detection rates as well in my experience.
     
  5. C.S.J

    C.S.J Massive Poster

    Joined:
    Oct 16, 2006
    Posts:
    5,029
    Location:
    this forum is biased!
    f-secure 2007

    55.5 mb ram usage
    13 processes
    full scan speed - 43 minutes 27 seconds
    boot up time - 1 minute 32 seconds
    F-secure Pc Performance Test Results = 265.0 (higher the better)
    ----------------------------
    bitdefender 10 internet security

    6.4 mb ram usage
    6 processes
    full scan speed - 13 minutes 8 seconds
    boot up time - 59 seconds
    Bitdefender10 suite Pc Performance Test Results = 269.7 (higher the better)
    ----------------------------
    Kaspersky Suite 6

    12.1 mb ram usage
    2 processes
    full scan speed - 28 minutes 40 seconds
    boot up time - 57.9 seconds
    Kaspersky Pc Performance Test Results = 270.8 (higher the better)
    ----------------------------
    Trend Pc-Cillin Internet security 2007

    142.8 mb ram usage
    7 processes
    full scan speed - 11 minutes 50 seconds
    boot up time - 1 minute 7 seconds
    Trend Pc Performance Test Results = 271.8 (higher the better)
    ----------------------------
    Avg Internet Security 7.5

    65 mb ram usage
    6 processes
    full scan speed - 15 minutes 58 seconds
    boot up time - 45.9 seconds
    AVG suite Pc Performance Test Results = 270.4 (higher the better)
    ----------------------------
    Nod32 Antivirus
    17.1 mb ram usage
    2 Processes
    ----------------------------
    Outpost 4

    6 mb ram usage
    1 Process
    ----------------------------
    Panda Internet Security 2007

    89.6 mb ram usage
    11 processes
    full scan speed - 7 minutes 51 seconds
    boot up time - 1 minute 37 seconds
    Panda Pc Performance Test Results = 271.0 (higher the better)
    -----------------------------
    Ca Internet Security Suite 2007

    114.6 mb ram usage
    13 processes
    full scan speed - 15 minutes 30 seconds
    boot up time - 1 minute 42 seconds
    CA suite Pc Performance Test Results = 268.4 (higher the better)
    -----------------------------
    Norton Internet Security 2007

    36.9 mb ram usage
    5 processes
    full scan speed - 14 minutes 21 seconds
    boot up time - 1 minute 13 seconds
    Norton Pc Performance Test Results = 269.2 (higher the better)


    ---------
    this was one i did a while ago, but it was the latest versions (unless there have been new versions in the past few months)

    the performance results were not as expected on my benchmark tests, but AVG booted the quickest and it also did 'feel the fastest' on pc performance.

    at this time, i never checked my dr.web, but im sure it would have whipped their ass for pc performance and 'up there with the best' in bootup times, but almost last in scan speed (i know this :) )
     
  6. SourMilk

    SourMilk Registered Member

    Joined:
    Mar 31, 2006
    Posts:
    630
    Location:
    Hawaii
    I don't own any licenses. :blink:

    *from the original post: "no changes were made to the configuration of the antiviruses - they were installed as downloaded."
     
  7. C.S.J

    C.S.J Massive Poster

    Joined:
    Oct 16, 2006
    Posts:
    5,029
    Location:
    this forum is biased!
    trials are free ;)
     
  8. Alphalutra1

    Alphalutra1 Registered Member

    Joined:
    Dec 17, 2005
    Posts:
    1,160
    Location:
    127.0.0.0/255.0.0.0
    Comparing RAM usage is pretty futile, since I don't have much knowledge on the subject, I will let vlk explain it better :p

    However, thank you for the tests. It would be interessing to know what settings you used on each av since that would dramatically affect the test. Also, what did you do to determine all of this data?

    Cheers,

    Alphalutra1
     
  9. SourMilk

    SourMilk Registered Member

    Joined:
    Mar 31, 2006
    Posts:
    630
    Location:
    Hawaii
    Nice work. I am too pragmatic for such detail and am thankful for those who can wring out the stats :) I didn't see Avast in your tests. I wonder what it would have done. I am not a fanboy of any particular AV but I am an avid gamer who is mostly a safe surfer. Doggone no cd patches might get me someday but until then I guess I will trade off detection for system performance.
     
  10. C.S.J

    C.S.J Massive Poster

    Joined:
    Oct 16, 2006
    Posts:
    5,029
    Location:
    this forum is biased!
    to be honest, at the time ... it was a suite thing so i didnt test my dr.web or avast.

    i through the processes and ram or outpost and nod, because they were popular at the time, i think they still are to be honest.
    ---------
    at the moment, im taking a risk with no real-time protection, got dr.web installed on vista just to use my licence *lol* even though it aint vista ready, everything works (to a point) apart from the real-time *spiderguard.

    update works through 'auto update' but not automatically.
    spam and email scanning works flawlessly
    ondemand scanning works flawlessly

    still waiting for the vista version, la la la la .
     
  11. SourMilk

    SourMilk Registered Member

    Joined:
    Mar 31, 2006
    Posts:
    630
    Location:
    Hawaii
    Hi. I'm not very scientific. I just downloaded the programs and installed them as mentioned on the orig. post. I did not change anything. No settings or configurations. While the AV's were installed with the guard on as was default on all the AV's I ran the CrystalMark benchmark. From all the different results like CPU, FPU, etc., I decided to use the overall system benchmark and the hard drive benchmark to show how much of a drag the AV guard would have on the total system and hard drive. I came up with numbers that I divided by the benchmarks of the standard installation without an antivirus installed to get the percentages.

    Like the others who have posted here, I am quite surprised that Avast did so well on my machine. Of course, it may not do so well on another computer.
    Here are the details of my computer:
    Asus P5nSLI, e6600, 2gb Kingston, 10k raptor HD, 8800 GTS, Windows XP (for now).

    Hope this answers most of the questions.
    Sourmilk out
     
  12. Alphalutra1

    Alphalutra1 Registered Member

    Joined:
    Dec 17, 2005
    Posts:
    1,160
    Location:
    127.0.0.0/255.0.0.0
    I think I need to read better, I missed the performance impact tester mentioned in your first post o_O

    I was just interested really in the settings, because I noticed that some avs, when cranked up to the max, don't perform very well. However, I guess since most users just keep default settings, your tests are more relevant to a greater majority.

    Thanks again for your effort.

    Cheers,

    Alphalutra1
     
  13. FRug

    FRug Registered Member

    Joined:
    Feb 7, 2006
    Posts:
    309
    I was kind of surprised by your measured HDD impact of AntiVir so i downloaded crystal mark 2004R2 and gave it a spin myself. My own results were even more surprising:

    The benchmarking values varied massively in successive runs, i only pick one value here, but the others were behaving similarly.
    without guard running:
    seq read speed first run 45mb/s
    seq read speed second run 40mb/s
    seq read speed third run 20mb/s
    seq read speed fourth run 46mb/s

    with guard running (all files)
    seq read speed first run 40mb/s
    seq read speed third run 21mb/s
    seq read speed second run 37mb/s
    seq read speed third run 47mb/s


    with guard running (smart ext)
    seq read speed first run 47mb/s
    seq read speed third run 48mb/s
    seq read speed second run 39mb/s
    seq read speed third run 48mb/s

    Looking at these results I am not sure what happens, what I am sure though is that crystal mark has some REAL issues to produce meaningful and stable results. With these variations in the result the test seems pretty damn random to me.
     
  14. The One

    The One Frequent Poster

    Joined:
    Mar 6, 2007
    Posts:
    246
    Nice test
    So I need to reconsider Trend or Panda?
     
  15. FRug

    FRug Registered Member

    Joined:
    Feb 7, 2006
    Posts:
    309
    gave it another spin, now some more detailed results, the first 3 are without active guard, the last two are with active guard and full detection settings + allfiles.

    Again the results vary too much under identical conditions for my taste.

    Code:
    No active guard:
    ================
    [ HDD ]              5946
               Read :   42.26 MB/s (  1690)
              Write :   42.41 MB/s (  1696)
     RandomRead512K :   22.66 MB/s (   906)
    RandomWrite512K :   25.92 MB/s (  1036)
     RandomRead 64K :    6.20 MB/s (   248)
    RandomWrite 64K :    9.26 MB/s (   370)
    
    [ HDD ]              5460
               Read :   31.37 MB/s (  1254)
              Write :   42.63 MB/s (  1705)
     RandomRead512K :   22.13 MB/s (   885)
    RandomWrite512K :   25.39 MB/s (  1015)
     RandomRead 64K :    5.96 MB/s (   238)
    RandomWrite 64K :    9.08 MB/s (   363)
    
    [ HDD ]              5468
               Read :   37.14 MB/s (  1485)
              Write :   39.06 MB/s (  1562)
     RandomRead512K :   21.09 MB/s (   843)
    RandomWrite512K :   25.00 MB/s (  1000)
     RandomRead 64K :    5.69 MB/s (   227)
    RandomWrite 64K :    8.79 MB/s (   351)
    
    [ HDD ]              4858
               Read :   32.17 MB/s (  1286)
              Write :   33.84 MB/s (  1353)
     RandomRead512K :   18.94 MB/s (   757)
    RandomWrite512K :   22.41 MB/s (   896)
     RandomRead 64K :    5.77 MB/s (   230)
    RandomWrite 64K :    8.41 MB/s (   336)
    
    
    Active guard:
    ================
    [ HDD ]              6093
               Read :   44.74 MB/s (  1789)
              Write :   43.64 MB/s (  1745)
     RandomRead512K :   23.20 MB/s (   928)
    RandomWrite512K :   26.03 MB/s (  1041)
     RandomRead 64K :    5.89 MB/s (   235)
    RandomWrite 64K :    8.88 MB/s (   355)
    
    [ HDD ]              5723
               Read :   38.45 MB/s (  1538)
              Write :   42.39 MB/s (  1695)
     RandomRead512K :   22.05 MB/s (   882)
    RandomWrite512K :   25.60 MB/s (  1024)
     RandomRead 64K :    5.82 MB/s (   232)
    RandomWrite 64K :    8.81 MB/s (   352)
     
  16. SourMilk

    SourMilk Registered Member

    Joined:
    Mar 31, 2006
    Posts:
    630
    Location:
    Hawaii
    If there is a problem with CrystalMark, which your numbers seem to indicate, then my tests posted earlier are flawed which may account for the Avast surprise. Maybe we should use another benchmark to see if it can shed a better light on the system performance impact by different AV's.

    Sourmilk out
     
  17. FRug

    FRug Registered Member

    Joined:
    Feb 7, 2006
    Posts:
    309
    SourMilk: Would be interesting to see whether your system gives such variable results as well.
     
  18. SourMilk

    SourMilk Registered Member

    Joined:
    Mar 31, 2006
    Posts:
    630
    Location:
    Hawaii
    I just ran the hard drive section of CrystalMark 3 times and received the following scores (no AV installed):
    8943
    8996
    8917

    It seems to be fairly consistent running in my machine. But, who really knows? The Shadow?:D

    SourMilk out
     
  19. vlk

    vlk AV Expert

    Joined:
    Dec 26, 2002
    Posts:
    618
    Just FYI, starting with the last version (4.7.942), avast has been optimized for the Intel Core 2 architecture. That's why it may be running faster on your E6600.

    Also, avast now has the "do not scan system libraries" setting that dramatically reduces the number of files scanned by the on-access scanner.
     
  20. lodore

    lodore Registered Member

    Joined:
    Jun 22, 2006
    Posts:
    9,006
    @chris how come you didnt put the scan time for nod32?
    lodore
     
  21. C.S.J

    C.S.J Massive Poster

    Joined:
    Oct 16, 2006
    Posts:
    5,029
    Location:
    this forum is biased!
    at the time lodore, i just couldnt be bothered, as i said... at the time, it was suites only.

    nods scan speed is damn fast though, we all know that.

    it scans my dual core on vista in just 8 mins on highest settings with everything enabled. *lol*
     
  22. SourMilk

    SourMilk Registered Member

    Joined:
    Mar 31, 2006
    Posts:
    630
    Location:
    Hawaii
    Thanks VLK. Mystery solved. I have always liked Avast. Even without heuristics, the guys involved get signatures out to the public faster than most AVs and have shown their stuff on the latest threats. Having Webshield, P2P, and other filters make Avast a definite steal for home users. Now with the decreased system performance impact with dual core cpus, gamers have another option besides NOD32.
    SourMilk out
     
  23. FRug

    FRug Registered Member

    Joined:
    Feb 7, 2006
    Posts:
    309
    I've just tried the benchmark on my notebook (previous test was on workstation), and there i also have approx. 25% variation of the results under the same conditions. Fastest result with active guard btw ;)
     
  24. lodore

    lodore Registered Member

    Joined:
    Jun 22, 2006
    Posts:
    9,006
    good could be useful for the computer im building soon
    i want an av that dont lagg my online game
    kav laggs it when it updates for a few secs
    lodore
     
Loading...
Thread Status:
Not open for further replies.