![]() The last page they posted on PCs being faster was taken down soon thereafter, amid community response about Premiere’s severe inferiority when compared with packages like Apple’s Final Cut Pro and Avid’s Xpress DV. If Apple can’t keep pace with the price reductions, we’ll be back in G3/G4-era again - too much cash from not enough performance.Īdobe corroborates that pcs are faster than G5. But that won’t last long, and eventually they’ll become commodity, and you’ll find them in $1500 Dells. Apple’s got a little bit of breathing room right now on the price front - Athlon 64’s don’t come cheap. What IBM needs to do is keep the clock-speed ramping up in line with x86 processors, and what Apple needs to do is find a way to sell them cheaply. Independent benchmarks (*never* trust benchmarks from the company that makes the machine!) show the G5 to be a decent bit faster clock-for-clock than a comparable Athlon64 machine. The G3 has hampered by crappy FPU performance, and the G4 was held back by a dated memory bus that kept AltiVec from showing its stuff. The G5 is Apple’s first decent CPU in many years. Adobe used to favor Macs over PCs, and I’m sure they took some flak for their preferred platform being slower than the competition. I think Adobe is just pissed at Apple for years of G3 and G4 crap. It will be at least a half decade before AMD64 processors can begin to be used to their fullest potential.Īs for the Athlon 64 being the fastest x86 chip, have any benchmark numbers to back that up? Unless I’m mistaken, the 3.2GHz P4 holds that title. Unlike the POWER architecture, which IBM has had over a decade to fine-tune their compiler for, AMD64 is a relatively new ISA, and compiler availablity is scant at best. ![]() On the AMD64 side, there is no highly optimized compiler waiting to be used. The full potential of the processor cannot be realized, however, until Apple defines a 64-bit ABI for OS X. The difference is that IBM will soon release a final version of their XL C compiler for OS X, at which point applications can be compiled in a truly G5 optimized manner. In fact, the Athlon64 enjoys that title, as well as fastest x86 chip.Ĭlock for clock, the G5 currently performs about as well as an Opteron when running G4-optimized code on the G5 and IA32 code on the Opteron. ![]() However, as I do work in scientific computing (mesoscale atmospheric modelling, to be precise, just check the domain of my address) I assure you, on our clusters we use Intel’s icc and ifc, not gcc. There are apologists who will make up claims that gcc is used for “scientific computing” which is who Apple allegedly was targeting the benchmarks at (although this is highly dubious). They ran SPEC CPU2000 on their competator’s systems for them, rather than going with the posted numbers, and compiled it with gcc, which crippled the performance. Well, their benchmarks certainly were dishonest.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |