29th June 2005, 7:02 PM
The article was removed, maybe Sony or MS's lawyers got to them... :D Seems pretty quick for that, though... but I can't think of another reason it'd dissapear.
Anyway, it said that the IBM PowerPC CPU is underpowered, essentially. This isn't new -- why do you think Apple gave up on IBM? Yeah. And remember that Apple said that they looked at the Cell, but it was lacking in some things they thought computers needed, so they didn't use it... anyway, it says that the Xenon has a 3-unit CPU (the main complaint here is that parallel processing won't happen for some time... the thing is designed to run multiple operations at once, but at first it'll just be doing one at a time, slowing things down quite a bit -- and it could take a while before people can run three or four operations at once like it's supposed to. This hurts speed, obviously.), all of which are general-use, versus the Cell with one core that has nine minicore things that do specific tasks. It then said that the uses of the minicore things (SPUs or something?) were, according to the devkits, pretty limited, so they couldn't completely make up for the fact that the PowerPC is a relatively weak CPU... then Anand proceeded to attack the relevance of the "flop" (operations per second count), saying that it's an irrelevant number with no real-world applications -- you can't just run math equasions on the CPU all the time, it's got other things it's got to do... so MS's "one Teraflop" vs. Sony's "2 teraflop" argument really doesn't matter. He then gave an example of the Pentium 4 vs. the Athalon 64 -- the P4 can do 15 billion flops while the Athalon 64 can do 8.5 billion, so the P4 is twice as good as the Athalon 64, right?
In the end, though, he did say that there was a saving grace... the GPUs. Both the PS3 and X360's graphics cards are said to be quite good...
Anyway, it said that the IBM PowerPC CPU is underpowered, essentially. This isn't new -- why do you think Apple gave up on IBM? Yeah. And remember that Apple said that they looked at the Cell, but it was lacking in some things they thought computers needed, so they didn't use it... anyway, it says that the Xenon has a 3-unit CPU (the main complaint here is that parallel processing won't happen for some time... the thing is designed to run multiple operations at once, but at first it'll just be doing one at a time, slowing things down quite a bit -- and it could take a while before people can run three or four operations at once like it's supposed to. This hurts speed, obviously.), all of which are general-use, versus the Cell with one core that has nine minicore things that do specific tasks. It then said that the uses of the minicore things (SPUs or something?) were, according to the devkits, pretty limited, so they couldn't completely make up for the fact that the PowerPC is a relatively weak CPU... then Anand proceeded to attack the relevance of the "flop" (operations per second count), saying that it's an irrelevant number with no real-world applications -- you can't just run math equasions on the CPU all the time, it's got other things it's got to do... so MS's "one Teraflop" vs. Sony's "2 teraflop" argument really doesn't matter. He then gave an example of the Pentium 4 vs. the Athalon 64 -- the P4 can do 15 billion flops while the Athalon 64 can do 8.5 billion, so the P4 is twice as good as the Athalon 64, right?

In the end, though, he did say that there was a saving grace... the GPUs. Both the PS3 and X360's graphics cards are said to be quite good...