Panda Software: Competitive Benchmark (Veritest, Dec 2006)

[Иван 290]

Нашел тут тест, он заказан Пандой, но тем не менее результаты остальных вендоров интересны. Если вкратце тест показывает насколько запущенная задача сканирования замедляет работу пользователя, которая имитируется различными тестами WorldBench 5.0

Panda Software: Competitive Benchmark

Test report prepared under contract from Panda Software

RETAIL PRODUCTS

MCAFEE

· McAfee Internet Security 2007 [7.0.329]

SYMANTEC

· Norton Internet Security 2007 [10.0.0.86]

PANDA SOFTWARE

· Panda Internet Security 2007 [11.00.02]

KASPERSKY INTERNET SECURTIY

· Kaspersky Internet Security [6.0.0.303]

TREND MICRO

· PC-cillin Internet Security 2007 [15.00.1433]

The purpose of the first test was to demonstrate the performance impact that the presence of an antivirus application has on a computer. Utilizing WorldBench 5.0 software from the PC World Test Center, VeriTest Engineers were able to obtain a baseline performance average of a computer with no antivirus software installed. Each individual antivirus application was then installed and tested for PC performance impact using WorldBench 5.0 software. The second test tracked the length of time necessary to scan a 1GB file set. This file set consisted of many

levels of folders and many different file types, including compressed files. Known infected files were included in order to assure that each product was scanning for viruses.

Testing methodology

VeriTest configured three identical Windows XP Professional workstations. Each workstation was a Dell Optiplex GX 620 with 3.4 GHz processor and 1 GB of RAM. An ATI Radeon 9250 video card was also added to each PC. VeriTest engineers installed and configured a fully patched installation of Windows XP Professional on each PC.

Impact on World Bench 5 Performance

WorldBench 5.0 was then installed and configured. Using Norton Ghost, a hard drive image was saved to a remote storage device so that the PC may be returned to a fresh baseline when needed.

Utilizing WorldBench 5.0 software, a “Full Run” test was executed on each PC. This test included three iterations wherein all tests would be completed three times to achieve an average result. WorldBench 5.0 software executed a series of tests to determine the time taken to complete each complex task. Upon completion of the “Full Run” series of testing, the results were exported. The PC was then returned to a fresh baseline state via restoration of drive image. These “Full Run” results provide the baseline performance statistics required to compare future test results against.

The WorldBench 5.0 benchmark software is very sensitive to other applications running in the same environment in which it operates. Due to this, some of the products being tested required configuration changes in order to suppress pop-up messages. These configuration changes are listed in Appendix B. Immediately after installation, each antivirus application was instructed to complete all available updates and patches via the Internet. Utilizing WorldBench 5.0, a “Full Run” test was executed twice, completing six iterations of each World Bench 5 application test. The results from each test were exported and averaged to produce the results shown in the Test Results.

WorldBench 5.0 is a tool for measuring a computers “real-world” performance capability. WorldBench 5.0 automatically executes functions within popular desktop applications. WorldBench 5.0 times the completion of these functions to produce its scoring results.

WorldBench 5.0 executes tests through automated functions using the following bundled software:

1. ACD Systems ACDSee PowerPack 5.0

2. Adobe Photoshop 7.0.1

3. Adobe Premier 6.5

4. Ahead Software Nero Express 6.0.0.3

5. Discreet 3ds Max 5.1 (DirectX)

6. Discreet 3ds max 5.1 (OpenGL)

7. Microsoft Office XP with SP-2

8. Microsoft Windows Media Encoder 9.0

9. Mozilla 1.4

10. Multitasking: Mozilla and Windows Media Encoder

11. Musicmatch Jukebox 7.10

12. Roxio VideoWave Movie Creator 1.5

13. WinZip Computing WinZip 8.1

[Dr.Golova 55]

Гм, странный тест. Если он заказной, то КАВ не должен идти в ровень с пандой. Можно же было подобрать тестовый набор чтоба медвед был явно впереди. Панда же явно тормознее шестерки на стандартном наборе пользователя. Что-то их ПР отдел халтурит.

[Сергей Ильин 1538]

Я думаю они правильно сделали, что по результатам они не односзначно самые быстрые. Во-первых, это выглядит более достоверно в глазах пользователей, во-вторых их цель была показать, что они быстрее Symantec, McAfee и Trend Micro. Тест эту задачу решает.