On Sunday 21 January 2007 14:40, Jan Engelhardt wrote: > >2. On my main workstation, glxgears got 10,000 frames/sec with the x86_64 > >install, and 7,000 frames/sec with the i386 install, a noticeable > > difference. > > I repeat it again: glxgears is *NOT* an appropriate benchmark. > And I doubt 3000 frames/sec really make a difference since humans > don't notice any improvement above 70 fps.
That is a specious argument Jan. Its faster in 64 bit. Period. By almost 30% What does it matter that a program has to test large numbers of events to accumulate a meaningful sample size? Are you saying that because a human could never even conceptualize adding 3.1459 to 239.732 in the time a computer could perform the operation a billion times we should just go back to the abacus? We don't need computers that are faster than us? You sit down at a computer that can muster at most 70fps running glxgears and see how happy you would be compared to using one that does 7000fps. Glxgears is as good a "quick measure" as anything else. Dismissing it because its faster than you can see is just silly. -- _____________________________________ John Andersen -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
