Originally posted by: dullard
As a longtime dual user, I think you are overestimating their performance boost. For programs not dual optimized you get nothing but the ability to run background programs trouble free. For programs that have dual capability, they ususally aren't perfectly optimized for dual processors (ie can you always divide a task perfectly such that both processors are 100% utilized - which when you include the OS requirements is not a 50/50 split and is quite difficult to do). Thus realistically most dual optimized programs will only net you ~30% gain. I use one program which is really well suited for dual processors and I only expect a 50%-80% gain when using it. So lets be very generous and give that PPC 997 a 50% boost for going dual. Thus using your 2.3 GHz HT figure, add 50% and you get 3.45 GHz. Basically it will then tie the 3.4 GHz P4 which should be released at about the same time. There won't be any crushing on average. Maybe some programs it will do quite well, but others it will do quite poorly, so on average I'd expect a close tie. Of course this all hinges on your own 2.3 GHz number and my generous 50% boost number.
Yeah, that's why I said "fastest
current P4". However, I do agree with your dual optimization comment. I'm really just talking best case scenario for specific apps. Some apps get a substantial boost from dual processors, and it's sometimes way more than 50%. But like you said, usually it is not. Either way, it's still gonna be fast though. As long as Macs and PCs run hard core number crunching apps in the same ballpark, Macheads will be happy.
By the time Macs are actually running their G5 1.8GHz, the P4 will be post-5GHz, 64-bit, and on a 1GHz+ FSB.
Not even close actually. The PPC 970 1.8 GHz is already out (sampling), and is slated for release in actual machines in 2003H2. They're 64-bit, and capable of a 900 MHz bus. This is confirmed for IBM, which will be using them in their Linux blade servers, and it is widely expected for Apple as well. (Remember that the PPC 970 has been Altivec'd, which basically makes it a lock that Apple will be using the chip. IBM traditionally ignored Altivec for chips they used themselves.)
Even Intel themselves said that we should only expect the P4 to be at "3.2 GHz or greater", 32-bit, on an 800 MHz bus, by the end of 2003. Even if
dullard is correct that they'll be near 3 and a half GHz, that's still a far cry from 5 GHz. We're not sure if even the 90 nm Prescott will even reach 5 GHz, and it's not even out yet.
Even if Apple were to go with say up to only 1.6 GHz chips in 2003, with 1.8 GHz chips not coming out until 2004, Prescott would likely only be up to 3.6 GHz by then. And IBM has stated that the PPC 970 should hit 2.5 GHz. (It's not clear if that 2.5 GHz is on 0.13 um or 0.09 um though.) See this press release from CeBIT in German:
"
The PowerPC Blade offers outstanding performance and is superior to Intel Blades for certain applications in the High Perfomance Computing Sector. It is ideal for very computing intensive applications, for example in the area of simulation like meterology or geological calculations. The PowerPC Blade integrates seamlessly into the IBM eServer BladeCenter architecture with all its software components. Power und Intel Blades can be mixed in a BladeCenter in any order depending on the software applications.
The new IBM PowerPC 970 is the heart of the PowerPC Blade. It is based on the 64-Bit Power 4 architecture which is also used in the processors of the IBM eServer pSeries. The 64-bit microprocessor:
- Offers full symmetrical multi-processing
- Has a high reliability (with parity L1, ECC L2 and parity checked system bus)
- Is manufactured in the latest 0,13 micrometer Copper/SOI CMOS technology
- Runs at frequences ranging from 1.8 GHz - 2.5 GHz"
A 2.5 GHz single processor would likely in the ballpark of a 3.2 GHz P4. And of course, a dual 2.5 would be quite nice.
Facebook
Twitter