Intel's Processor Plans

[arn]

This CNet article on Intel's possible backup plans to it's Itanium include a possible "32-64" bit version of the Pentium. This hypothetical Pentium would be in place if the Itanium (new architechture) were never to gain acceptace. Of interest regarding the architechture switch:

Few companies have ever successfully switched from one architecture to another. Intel, in fact, attempted to switch from its "x86" architecture, the architecture behind its microprocessors, in the early '80s. Many of the company's best designers were put on the Sierra project, a chip based around a budding new architecture that would succeed the 286 chip, according to Brookwood. In the end, however, the company went with its plan B, an extension of the 286 called the 386. Pat Gelsinger, one of the leaders on the 386 project, has since become Intel's CTO.

Few companies indeed... with Apple being one (if not the only).

 

[arn]

regarding relevancy

Before anyone starts posting that this isn't a Mac/Apple story...

I'll just preempt it before it starts... information on the industry as a whole is relevant in that it relates to Apple's future. While every non-Apple rumor may not be relevant, an occasional key one on the rest of the industry provides for well rounded 'apple-rumor-follower'

arn

 

[irmongoose]

Hmm.. so... the Pentium is becoming like the Itanium?

Well... no more megahertz myth then, huh... since the Itanium is 800MHz but its faster dan Pentium... just like the G4.



irmongoose

 

[pc_convert?]

I think it's a pretty safe bet Itanium will fail. Intel was forced to change from the x86 architecure to break into the high-end server market in order to compete with SPARC, and Alpha. AMD on the other hand targets the home user, so to get more 'omph' out of the x86 architecture they naturally decided to extend it to a 64bit version (x86-64).

While Itanium costs $1000's dollars per chip becuase its targeted at high end servers, and uses around 130W of power. AMD's Hammer chips should cost less because they are aimed at the home market.

Now if the G5 is ready, Apple could really make some inroads into the server and home markets by beating AMD to the market with the first 64-bit home computer. Remeber the press when PC's broke the 1GHz barrier, think of the press if Apple could be first to the market with a 64bit computer.

 

[Unregistered]

Apple isn't the only one. Back in 1987, Acorn switched from using the 6502 processor in it's "BBC" range of computers to their in-house developed RISC Micro called (then) the Acorn RISC Machine.

The name eventually changed to ARM...

Jim

 

[mac15]

Somebody got hold of the force
the future is near

 

[pc_convert?]

Just to expalin my "I think it's a pretty safe bet Itanium will fail" statement.

AMD's decision to extend x86 has forced Intel into having to have a contingency with it's own 64bit extension. Home users will buy the Hammer CPU's because it's affordable 64bit computing, that doesn't require all new software because it backwards compatible.

Itanium is targeted at the server market. It costs a fortune and to get speed improvements you need to re-invest in Itanium software. Yes Itanium runs x86 code, but very very slowly - which is just great for a server!

So AMD sells it's x86-64 CPU's by the bucket load, server vendors think why buy Itanium - involving huge hardware and software investment when you can buy 2,4,8-way x86-64 CPU's at a resonable hardware cost, using the same software.

Intels knows this will happen so has to turn on the 64bit extensions in the newer Pentiums, so the server market has 3 64bit choices

1) Low Cost 64bit AMD.
2) Low Cost 64bit Intel.
3) High Cost Intel Itanium.

Given the current economic climate option 3 doesn't look good. Intel writes off a $billion dollar+ investment.

Apple then suceeds in world domination, world peace is achieved, and humanity enters a golden age.

erm that last part is a dream....

 

[zephyrus]

64bit... Who cares? Be it Mac or PC (x86 or PPC) a 64bit processor does NOT mean that it is faster. In fact, it may end up being slightly slower than a 32bit processor at the same clock speed. So my question is... Why would a home user care? It does make perfect sense in the server market where you have 64bit apps that need the expanded throughput.

Now, the G5 having multiple cores on one dye... Now THAT is cool. A dual processor G5 would be like a quad processor G4 but with better throughput between the cores. Now THAT is cool.

 

[me hate windows]

screw intel. I hope AMD kills those @sses. Apple rules

 

[dieter]

Apple isn't the only one. When Sun Microsystems switched from their 68k-based Sun-3 series to the SPARC processor in the early 90s (?), they introduced a completely different architecture, which has been successful so far.

 

[blindman858]

AMD

Did u know that AMD could be driven bankrupt with itanium!!!!! lol They are in the works with something called "the hammer" which is supposed to be a itanium competitor. I for one find amd chips much better than intels what do u think?

 

[jaykk]

Mac & AMD

Here is an interesting article from Inquirer
http://www.theinquirer.net/30010215.htm

 

[Onyxx]

from what people that have actually used Itanium machines (ibm workstations) their opions can basically be summed up in one sentence "Itaniums are blazing fast but they suck." their reason is the poor way that the chip is used by software. what good is the world's fastest chip if there is nothing around to use on it? Right now windows software has to go through a sort of x86 transcoder/emulator thingamobober in order for it to work. this makes it slower than it could be.
The truth is that even if Intel made 64 bit processors it wouldn't matter. Mac software can turn on a dime and will be converted to a 64 bit native version much faster than windows software (granted not all companies are as fast as some ex. Adobe). How long would it take microsoft to recompile a 64 bit version of windblows xp? that thing is monsterous and the code for it looks like a copy of war and peace that has been through a paper shreader, glued back together at random and rewriten in C. shiver
Yes the unix world will rejoice. But the thruth is that there are just not a lot of unix people out there and i doubt intell would risk scaring off the consumer market with halved clock speeds (or more) just to get a 64 bit chip out.

just another ¢2

 

[tadpole]

itanium, why it simply sucks

well, actually, the unix world is huge, growing and generally thriving. why itanium sucks, well it has to fight: Alpha, UltraSparcIII, SGI MIPS, HP, IBM Power4, and hopefully G5 for it to become an accepted server chip. intel has to wage an uphill battle with well established unix box chips with PROVEN track records and highly tuned software in order to crack the high performance market. meanwhile AMD will clean clocks on the home front with their Hammer series. Intel/Apple/AMD excell on the low end server, and desktop client markets. In academia anyhow at my campus and at the better funded Big Ten and Pac Ten campuses, all the backbone equipment, all the research infrastructure, all the main servers are unix. Mostly a gimish of IBM RS/6000 based machines and Sun Sparcs. My department (biochemistry and molecular biology) would rather sell back their phd's then run something less than a proven unix box, sometimes we run linux on old hardware. OSX and linux are nice introductions to the real world of unix.

I've been flamed so many times for thinking the world of unix, some guy once told me that in the real world only wintel machines rule. well, lemme say this- if you're GM and you're going to do modelling of a vehicle's crash performance, are you going to run that on a wintel or an SGI Onyx? If PC's are so great (Macs are also PC's) then why are all the engineers in the cellular group at Motorola Boyton Beach using Sun Sparcs?


What would be nice, a G5 with full on 64bit ISA, full speed L2 cache at least 4 megs in size, a Unix98 compliant version of OSX- that's right, make it a real unix, not a unix-like operating environment, 64bit 66mhz pci at the very least, do away with the stupid 64bit 33mhz pci, DDR333 ECC ram- and Apple needs to stop shipping CL3 modules, CL2 modules give a small but nonetheless quantifiable performance increase- 10 to 15% speed enhancement, SCSI160 F-AL drives, and WORKSTATION class Graphics. See, this way, all the stuff already written for the Unix bad boys could simply be recompiled for Mac without being recoded. Imagine this, Mac's becoming a major workstation player, now how nice would it be for the consumer that they are also benefiting from research done to produce the workstations!