AV-Comparatives Performance Test October 2019

[polishpatriot]

Could you give us more information about it?

Just compare memory usage between Antimalware Service and say, for example, Bitdefender's usage or Kaspersky's.

And as you well know, WD doesn't handle large file system stores very well - a long standing problem. That can only be fixed by proper cache usage and optimization.

 

[Andy Ful]

Just compare memory usage between Antimalware Service and say, for example, Bitdefender's usage or Kaspersky's.
...

Are you sure?
Memory usage (approximately):
Idle: Defender 120 MB (190 on my computer), BitDefender 640 MB, Kaspersky 320 MB.
Scanning: Defender 525 MB , BitDefender 855 MB, Kaspersky 435 MB.
Updating: Defender 275 MB, BitDefender 690 MB, Kaspersky 495 MB.

And as you well know, WD doesn't handle large file system stores very well - a long standing problem. That can only be fixed by proper cache usage and optimization.

I do not think that this can have an impact on the performance in the home environment. This could be an issue on Windows Server with disabled auto-exclusions:

Microsoft Defender Antivirus exclusions on Windows Server - Microsoft Defender for Endpoint

Windows Server includes automatic exclusions, based on server role. You can also add custom exclusions.

docs.microsoft.com

I think the WD caching is an issue because it is cleared when closing the system and not because it is poor. Other AVs probably do not clear the cache in this way. But, If the cache is not cleared frequently, then the AV has to apply a method to check the cached files frequently against new signatures. If I correctly recall, then BitDefender free constantly checks the files on disk (with low priority and low CPU usage).

 

[polishpatriot]

Are you sure?
Memory usage (approximately):
Idle: Defender 120 MB (190 on my computer), BitDefender 640 MB, Kaspersky 320 MB.
Scanning: Defender 525 MB , BitDefender 855 MB, Kaspersky 435 MB.
Updating: Defender 275 MB, BitDefender 690 MB, Kaspersky 495 MB.

Higher memory usage of Bitdefender and Kaspersky is much better optimization.

I do not think that this can have an impact on the performance in the home environment. This could be an issue on Windows Server with disabled auto-exclusions:

WD cannot handle folders or drives with large number of files. It is because it uses sub-standard optimization (cache). It's the reason WD has a melt-down when you open a file containing hundreds or thousands of files, but that folder has been attached to the file system without ever rebooting the system.

 

[Andy Ful]

polishpatriot,
I think that we talk about two different things.
The problems which are reported by home users, like a delay with opening the folder with hundreds of executables (first time after reboot), are related to the fact that WD treats such a folder as if it was never opened and never checked (cached) by WD.
I performed a simple test:

1. Disable WD real-time protection after reboot and open such a folder in Explorer (no delay).
2. Close Explorer and enable WD real-time protection.
3. Open this folder again = no delay!!!

In fact, I am not sure if WD performs any file caching while opening folders or simply checks if the folder was accessed in the current Windows session. If the folder was already opened (with file checking or without), then the second time the files are not checked.

Anyway, it may be that such behavior is intended by MS due to the known problems with file caching.

 

[polishpatriot]

polishpatriot,
I think that we talk about two different things.
The problems which are reported by home users, like a delay with opening the folder with hundreds of executables (first time after reboot), are related to the fact that WD treats such a folder as if it was never opened and never checked (cached) by WD.
I performed a simple test:

1. Disable WD real-time protection after reboot and open such a folder in Explorer (no delay).
2. Close Explorer and enable WD real-time protection.
3. Open this folder again = no delay!!!

In fact, I am not sure if WD performs any file caching while opening folders or simply checks if the folder was accessed in the current Windows session. If the folder was already opened (with file checking or without), then the second time the files are not checked.

I see the non-cache delay when opening folders with lots of files - such as pCloud or OneDrive.

To me it, it is based upon observation. It can be said to be anecdotal evidence. Some users seem to have a lot more problem with it than others.

I am pretty sure WD does not use the equivalent of iSwift and iChecker technologies (Kaspersky). WD is using 2010 technology and methods. And like I pointed out indirectly, Microsoft subcontracts a lot of Windows Defender out to India and Eastern Euro shops. Those shops can only do what Microsoft specifies in the contracts. Otherwise I'm sure that those shops would make WD a better product.

I think MS deliberately leaves the "large folder" delay in W10 just to annoy power-users. Average home user doesn't have 250+ object-stuffed folders and drives.

 

[Andy Ful]

...
Some users seem to have a lot more problem with it than others.
...

I think so. It is not a problem in my case with folders on SSD. To see this delay I had to move the 2GB folder with many executables to non-SSD disk.

I am pretty sure WD does not use the equivalent of iSwift and iChecker technologies (Kaspersky). WD is using 2010 technology and methods.

After my test from the previous post, I am not sure now if any "real caching" of clean files is done, while opening the folders.

 

[polishpatriot]

After my test from the previous post, I am not sure now if any "real caching" on clean files is done, while opening the folders.

I just let it go and try to not think so hard. 5 seconds won't extend my life, but it sure might save my computer's.

 

[[correlate]]

15 Corporate Solutions Put to an Endurance Test

 

[Aggravatorx]

I have to say i do not use windows defender but this is already built into the system i never felt any slowness on any of the computers i own
it was actually the fastest think about it your downloading a big security program which most are huge installs adding all these protection
modules check task manager its slowing you down even eset in time effects the system.but as a MalwareTips user said todays computers with
the lateset processors and memory 16 gb more and ssd hard drives any one of these programs will run fine unless your still using a netbook

 

[RKRN3]

Avast was doing everything well, except the data selling, but no one has yet been able to pinpoint anyone from the data.

 

[manchesterunited]

AV lab tests are designed to show that a product meets a minimum set of standards.

AV lab tests do NOT go on fishing expeditions to figure out how a product will perform under the most adverse conditions at maximum settings.

That second testing scenario returns results in the 40 to 60 % range, sometimes less.

 

[Handsome Recluse]

There are many performance tests that show it and there is no contradiction between different tests because they usually test different performance areas.

Does simply looking at averages even predict anything in performance tests?

 

[Andy Ful]

Does simply looking at averages even predict anything in performance tests?

Probably not. Every AV Lab calculates different types of averages (with different weights) and can take into account different parameters.