ASUS T9 Notebook

Performance - Content Creation

The SYSMark 2001 benchmark also includes a Internet Content Creation part that uses typical Internet content creation tools such as Photoshop, Premier, Dreamweaver, and the like to rate how a system performs when performing these fairly common content creation tasks. How did the notebook do in these tests?

Content Creation Performance Internet Content Creation SYSMark 2001 Desktop 900MHz Desktop 800MHz ASUS T9400 900MHz Desktop 733MHz Desktop 600MHz ASUS T9400 700MHz Desktop 533MHz Desktop 466MHz 101 93 87 86 71 69 63 56 | 0 | 20 | 40 | 61 | 81 | 101 | 121

On AC power at the 900MHz clock speed the ASUS T9 was able to perform at about the same speed as a desktop system running at 733MHz. This left the notebook running about 14% slower than the desktop 900MHz testbed we set up. When the notebook kicked into 700MHz mode when on battery power it performed like a desktop 600MHz CPU, placing it about 20% behind the desktop 733MHz system. Again, we can attribute the notebook's sub-desktop performance to its system bottlenecks created by the hard drive, chipset, video solution, ect.

The last performance test we subjected the ASUS T9 to was Ziff Davis Media's Content Creation Winstone 2002. As mentioned before, the benchmarks we selected stress the notebook in the ways it would be stressed in a real life situation: performing simultaneous Internet content creation and office type tasks. If you forget exactly what the new Content Creation Winstone consists of, check here for a reminder.

Content Creation Performance Content Creation Winstone 2002 Desktop 900MHz Desktop 800MHz Desktop 733MHz Desktop 600MHz Desktop 533MHz ASUS T9400 900MHz Desktop 466MHz ASUS T9400 700MHz 15.8 15.6 14.5 13.3 11.9 11.7 10.9 10.2 | 0 | 3 | 6 | 9 | 13 | 16 | 19

Content Creation Winstone 2002 completely dominated the ASUS T9. Unlike before where we saw the notebook on AC power performing within 14% of its 900MHz desktop counterpart, the system fell 26% short in the Content Creation Winstone 2002 tests. This left the notebook performing more like a 533MHz desktop than a 900MHz one. When the power cord was severed, the system began to perform like a 466MHz desktop, falling well short of what we would expect from at 700MHz notebook. Why did the notebook perform so poorly in Content Creation Winstone 2002 compared to how it did in SYSMark 2001?

We suspected that perhaps the hard drive in the notebook was setting us far back. Compared to desktop hard drives, notebook hard drives are much slower and are typically pegged as the largest bottleneck in notebook systems. To test this theory out, we yanked the IBM Travelstar 30GN from the T9 and hooked it up to our desktop testbed via an 2.5" laptop hard drive to desktop IDE adapter. We installed Windows XP Professional and the required drivers and went to testing. Below is what we found.

Content Creation Performance Content Creation Winstone 2002 Desktop 900MHz Desktop 800MHz Desktop 733MHz Desktop 900MHz Notebook Drive Desktop 600MHz Desktop 533MHz ASUS T9400 900MHz Desktop 466MHz ASUS T9400 700MHz 15.8 15.6 14.5 13.5 13.3 11.9 11.7 10.9 10.2 | 0 | 3 | 6 | 9 | 13 | 16 | 19

Switching from the desktop Deskstar IBM drive we were using before to the IBM Travelstar allowed us to find the culprit of the lackluster Content Creation Winstone 2002 performance. Simply switching from the desktop hard drive to the notebook one resulted in a 15% performance decrease on the desktop system running at 900MHz. This instantly brought down the performance of the 900MHz desktop to the 600MHz level. It seems that we have found at least one major bottleneck in the ASUS T9: a bottleneck that will be present in almost every notebook but will vary depending on the notebook drive used in the system.

To see how much performance fell in the SYSMark 2001 benchmarks when the desktop hard drive was replaced with the notebook one we also ran the SYSMark tests on the setup described above. We found, surprisingly, that each SYSMark 2001 score remained the same regardless of the hard drive used. This suggests that Content Creation Winstone 2002 is much more disk intensive than SYSMark 2001; a hypothesis supported by the fact that the notebook system did much better in the SYSMark 2001 tests than it did in the Content Creation Winstone 2002 tests and proven by figures 10 and 12 in SYSMark2002's whitepaper overview which includes numbers for SYSMark2001 I/O requirements.