Walter Shillington is an avid collector of mechanical timepieces. While he focuses mainly on watches, Walter also reviews household items.
A Review of Futuremark PCMARK 7
PCMark7 is designed to benchmark the performance of computer systems running Windows 7. It includes seven PC tests covering storage, computation, image and video manipulation, web browsing and gaming.
This software is intended for the Information Technology professional and the computer enthusiast.
PURCHASE AND INSTALLATION
I downloaded Futuremark’s free version of PCMark7. Installation proved trouble-free.
WHAT IT DOES
The PCMark tests include Video playback and transcoding, System storage, Graphics, Image manipulation, Web browsing and Data decryption.
If you purchase the Advanced Edition, DirectX10 and DirectX11 tests can be performed. Specialized tests designed to assess netbooks as well as entertainment, creative, productive and storage based systems are available. For a list of features available on the Advanced and Professional Edition, please check the Futuremark website.
THE TEST DRIVE
I loaded PCMark7. As it began, Acronis non-stop backup paused. At the time I thought PCMark had stopped this operation to avoid diverting the computer’s resources. This proved not to be the case. Non-stop backup was not paused during further runs of PCMark.
The first video test displayed a busy city intersection - many lanes of moving cars, police on horseback and pedestrians crossing the street. This video, like all of the tests, repeated three times and the final score was averaged. The video playback test determines if the hardware is powerful enough to playback video at its intended rate without glitches. Video transcoding and downscaling are workloads that examine the capability of the hardware by changing resolution, framerate and bitrate of the video.
System storage tests isolate the performance of storage devices. The storage gaming test involves the creation of a character in World of Warcraft .
The performance of DirectX 9 graphics are examined using an animation of two fireflies making their way through a jungle.
The image manipulation test simulates everyday photographic procedures such as color correction, flip, rotate and stretch. The results are timed.
The Importing pictures storage test simulates the insertion of a thumb-drive containing 68 images. These images are imported to Windows live Photo Gallery. The test records activity on the system drive and measures the time.
The Web browsing test involves measuring the time required to open tabs in Windows Explorer. This test confused me, at first, because it worked even with the computer disconnected from the internet. Futuremark did not cheat - a copy of Windows Explorer is embedded into the program.
The Windows Defender storage test uses a trace of Windows Defender. A measurement is made of the time required to perform a quick scan on the storage device under test.
I ran this program six times. The final two attempts were with every non-Windows service terminated and the internet disconnected.
PCMark7 correctly identified the computer’s AMD Athlon X2 5400B processor. The Processor Clock is rated for 2800MHz. During the tests PCMark identified the clock speed at six different levels, ranging from 997MHx to 2793MHz. It correctly identified my ATI Radeon 3100 graphics, operating system, motherboard and memory.
The overall PCmark number is calculated from the set of test scores. Over my six benchmarks the score ranged from 1007 to 1102. By my calculations the scores varied by 8.6 percent. Below are the main test results
- Video playback results varied by twelve percent.
- Video transcoding – downscaling results varied by 11.1 percent
- System storage – gaming results varied by 14.7 percent
- DirectX 9 graphics results varied by 2 percent.
- Image manipulation results varied by 7.9 percent.
- System storage – importing pictures results varied by 7.4 percent.
- Web browsing results varied by 20.6 percent.
- Web browsing and decrypting/Data decrypting results varied by 10.2 percent.
- System storage – Windows Defender results varied by 9.3 percent.
I’ve worked as an electronic technician in the navy. When a transmitter is repaired a series of checks are performed to insure it meets specifications. Every three months these tests are repeated and checked against the prior benchmark. This helps to identify gradual degradations before they became a concern.
The test equipment is periodically sent to Marconi for calibration. At the calibration shop expensive and extremely precise instruments are used to verify the accuracy of the test equipment. In the case of transmitting equipment, there is wiggle room – specifications for a five volt test point might read 4.69V to 5.31V. The transmitter would work well if the output voltage was within that range. For test equipment, however, there is little wiggle room.
In my first benchmark PCMark7 clocked my computer’s processor at 997MHz. I repeated the benchmark. PcMark7 come up with a figure of 2791Mhz. I duplicated the process several times, striving to achieve repeatable results. Sadly, the readings varied widely. I’m uncertain as to the cause of these results. Perhaps these individual clock readings provide valuable information to an engineer. Maybe they are simply wrong. In any case this shook my faith in the accuracy of the software.
PCMark7 is a huge program, in part because traces of other software have been imbedded. The tests simulate real-world workloads and situations. Except for the general PCMark score, real figures are used. Each test is performed three times and, in my opinion, the results should give an accurate picture of how well the subject computer performs.
But it doesn’t. Over the course of six benchmarks the DirectX 9 score varied two percent. I find that acceptable. The Web browsing test results, on the other hand, varied 20.6 percent.
This, unfortunately, might be norm when benchmarking computer systems. Even if nonessential software is terminated, certain Windows processes can start and stop while testing is under way. I will be examining other benchmarking programs over the next few months. Once I figure out how Futuremark’s product compares, I will update this article.
This links to the Futuremark website
World leader in 3D, PC and mobile performance benchmarks. 3DMark and PCMark software estimate hardware performance in current and future games and applications.
wshillington on April 20, 2012:
Thanks, JAM671. You might be right. It certainly was frustrating to find this reading was not consistent.
JAM671 on April 20, 2012:
The results for your processor clock speed probably varied because of power management- on my system, the multiplier is automatically reduced at certain times to reduce power consumption. I can disable this feature in my motherboard's BIOS, and I think there also may be an option in Windows advanced power settings.