Als PDF hier: http://homepage.mac.com/WebObjects/FileSharing.woa/wa/default?user= danny_gasperini&templatefn=FileSharing6.html&xmlfn=TKDocument.6.x ml&sitefn=RootSite.xml&aff=consumer&cty=US&lang=en hier Ausschnitte: First impressions "THE SCREEN: The display is very bright and clear with good white point and gamma as the default settings. The brightness is on a direct par with the Titanium PowerBook G4, and slightly better than the iBook. Viewing angles, both vertical and horizontal, are not quite as good as the Titanium PowerBook. "THE KEYBOARD: In a word, fantastic. It is super smooth and quiet to use. I thought the previous PowerBook Titanium keyboards were nice, but these things are amazing. I am still not 100% sure about the material used in the construction of the keys themselves. They are probably a stenciled clear plastic, as the letters/numbers themselves are no longer painted. The 17" PowerBook keyboard is back-lit and these clear letters enable the light to transmit through "CONSTRUCTION: The quality of the assembly and fit and finish is impeccable. Apple has made a name for itself as the Mercedes of the computer industry in terms of refinement and this machine is no exception. Each of the screws used are uncannily flush with the housing, there are no uneven gaps when the lid is closed, and the batterys exterior surface is made of the same material as the case, unlike the Titanium - a nice touch. The clutch assembly used in the hinge is very fluid, and there is none of the "stickiness" that plagued the initial iBook and Titanium enclosures. "The benchmarks reveal some quite interesting information. For once, the Xbench data seems to be consistent with the real world benchmarks in most cases. They show that the CPU tests for the PowerMac and the Titanium are the strongest, which is probably due to the 1Mb of level 3 cache, and the fact that the PowerMac is running a dual processor configuration. The memory tests are strong across the board with the exception of the iBook, most likely constrained by its PC100 SDRAM and 100MHz system bus, while the other units are all running on a 133MHz bus with DDR RAM in the new 12 PowerBook and the PowerMac. This bodes well for the 17 PowerBook when it is introduced which runs a 167MHz bus and utilizes DDR 2700 SO DIMMs, and preliminary tests shown here (http://www.barefeats.com/pb17.html) would seem to reinforce this. The Xbench graphic tests are most intriguing. Quartz performance is strongest on the Titanium and the PowerMac systems, and given they have graphics chips based around the ATI Radeon Mobility 9000 and the nVidia GeForce 4 440MX respectively this is not suprising. I also suspect that they gain advantages from the L3 cache architecture they both share. What is suprising, however, is the strength of the PowerBook 12 nVidia GeForce 4 420Go performance in OpenGL, almost equalling the PowerMac, and substantially stronger than the Radeon 9000 Mobility in the Titanium. The first of the real application benchmarks reveals a clear victory to the nVidia based PowerMac in Quake 3, which is a game also optimised to take advantage of dual processors under OSX. The 12 PowerBook and the Titanium, however, share exactly the same score, even after running the test 3 times, and averaging the results. The Radeon 7500 based iBook posts a credible score in this test. Return to Castle Wolfenstein reveals another victory to the PowerMac, however, once again, the results are very close for the Titanium and the 12 PowerBook, with the iBook noticeably slower. RTCW seems (from past experience) to also be quite processor dependent which makes the performance of the 12 with no L3 cache even more satisfactory. "Jedi Knight II is an almost identical replica of the RTCW results, except that the iBook performs more strongly here. JK2 is another game built on the Quake 3 engine, except that is a more highly optimised engine and is much more texture intensive (like RTCW), which is why the frame rates are substantially lower than for Quake 3 Arena. The next test is a game that is also MP aware, and hence the PowerMac easily takes the crown in Nascar. The Titanium gains a slightly larger advantage over the 12 PowerBook than in other tests, but once again it is still fairly close, and Nascar is also a very processor bound application which once again raises the issue of a lack of an L3 cache. Finder operations on the new 12 PowerBook are very encouraging, with the Titanium and the PowerMac all being very close to each other. The iBook suffers quite markedly in this test as a result of its G3 architecture lacking the optimisations for the velocity engine and multithreading under OSX. Hard disk speeds show that while the PowerMac HD is substantially faster (thanks to its UATA/100 bus and 7200rpm UATA/100 drive), the other systems are all virtually identical. Even though the new PowerBook 12 also runs an UATA/100 bus, it shows no advantage, and is probably constrained by the maximum rotation speed of the drive itself. To conclude, the 12 PowerBook appears to perform very well in modern applications such as the games tested here. Considering that it runs no L3 cache, and utilises the nVidia 420 GPU that has only half the peak memory bandwidth of the Radeon 9000M and the nVidia GeForce4 440 GPU used in the 17 PowerBook (also with double the VRAM), it has managed to post some credible scores. The 17 PowerBook should show some substantial performance gains over the 12 PowerBook, and current Titanium 867 and 1GHz units, based on extrapolating these results, and taking into consideration the other major architectural improvements it sports.