PassMark vs SysMark

[rnockels]

There appears to be a fairly large discrepancy between PassMark and SysMark scores. To my knowledge, SysMark evaluates cpu performance based on real world applications rather than a synthetic test. Which should I use when making a choice for a business PC?

[David (PassMark)]

The last full release of Sysmark seems to be from 4 years ago (Sysmark 2004). They had a 2007 preview, but never seem to have made a final release.

Their 2007 preview DOESN'T support Window 2000 nor XP 64bit, nor Vista 64bit. it also only supported on English versions of Windows. it also required 30GB of disk space and 1Gb of RAM.

PerformanceTest requires 10MB of disk space (3000 times less than sysmark), and we support Windows 2000, XP & Vista (32bit and 64bit). Our tests include native 64bit code to test 64bit CPUs.

Sysmark contain custom version of Adobe and Microsoft applications (like Excel 2003 and Powerpoint 2003). But, according to the Sysmark develoeprs, "If you have previous or current versions of these applications already installed, they will conflict and possibly corrupt your applications, your benchmark scores, or your entire operating system."

So it it true to say that Sysmark uses real world applications to calculate a benchmark result. But that doesn't imply it actually places a high load on the CPU. For many of their tests the disk speed is probalby the bottleneck (this is just an assumption, as we didn't spend the $500 to buy their benchmark tool).

When was the last time you saw Microsoft Word 2003 max out all four cores in the quad core CPU?

If you are only running Microsoft Office on your PC (and have $500 to spare), then Sysmark would be an appropriate benchmark tool.

If you want to know about the maximum hardware capability, or gaming performance, or database performance, etc, then sysmark isn't a good choice.

[wonderwrench]

It's also nice that Performance test will run off a flash drive. No bench tool is the final word on performance but a good one like PT can be used as a quick check to see if a given PC is performing as expected compared to other similar systems. It's best use is to compare changes made on a specific system.
Like a video driver or video card change. Ram timing or ram change etc. After you use PT for a while you get a good feel as what to expect from a given system setup. I run PT on every system I build or work on as a final check to make sure all is well. I save the results so if I see the system later I can compare the results to see any changes. This is very useful as PT can point you to the area causing possible performance problems. Oh and by the way I only build/repair PC's for fun not profit.