> - What about this represents a work load that (you claim)
> won't show up until 5 years in the future?
> - What about this documentation does not (as you claim)
> justify the changes in the usage model?
> - What about this benchmark is so different than how
> 99.99% of the populace (that you claim) doesn't work in
> this way?
There's not enough information here. Does the system run ALL of these benchmarks at the same time? Is WME done in the background? If so, it may take up a majority of resources, as something like a compressed video or audio file is a major hog, much more so than a Word or Excel macro. If you're testing WME at the same time as the Office macros, then maybe you're only really testing WME performance, or maybe WME has a very strongly disproportionate effect on performance.
How exactly are they testing Photoshop? Are they using the most commonly used filters and actions, or are they stressing performance with uncommon filters and actions? There's been a lot of grumbling about the percentage of SSE-enabled filters in tests like these being *a lot* higher than the percentage of SSE-enabled filters actually used by real people who are mucking around with the images. Is that the case? It is BapCo's responsibility to clarify in deep detail what they're doing. This descriptions you've given me barely scratch the surface of that responsibility. They're getting paid a lot of money for these testing suites, so they had better be obligated to publicly prove their products' worth!
> - What about this situation (as you claim) artificially
> favors Pentium 4 over Pentium III?
I didn't claim that it artificially favours P4 over PIII (well, if I did, then I worded something wrong, or lied, or something<g>). I said that it might. And, well, it's obvious that the P4 did get a rather sizable boost from the prior version of SysMark to the current version of SysMark. Why is this? If it's because of "future behavior", BapCo has yet to really justify it in detail. |
