AMD 64bit vs Intel 3.2

[reubenb]

What is better?
AMD 64bit 2GHZ CPU or Intel 3.2GHZ "EXTREME" CPU?

[liorean]

That's not the right question. What do you want to do with it? Intel IA32 are far better for running current 32-bit applications, games, and such. The hammers (x86-64) on the other hand are far better at heavyweight calculations, 64-bit and 128-bit multimedia handling, etc.

Also, the Athlon64 family isnt' that much better than the same number of Hertz pentium at 32-bit, while they are about 1.5 times their speed at 64-bit compared to the Itanium2.

[reubenb]

ok
so i want the bst computer for performance of heavycpu programs..
what do u recommend

[liorean]

Again, wrong question. The IA32 is better at 32-bit, the x86-64 better at 64-bit. Thus, a 32-bit heavy CPU program would run faster on the Intel than the AMD, and conversely a 64-bit heavy CPU program would run very much faster on the AMD than the Intel (given the 64-bit program is coded in a way that even allows the Intel to handle it).

The question is rather if you want to run current programs for 32-bit systems or whether you want to run future programs for 64-bit systems. There's a possibility that software manufacturers will create programs that use 64-bit if they can, 32-bit if not. Drawing a parallel to G3-G5 on the mac, you can see Adobe doing such a thing for instance.

[surfie]

It would help a bit if you were more specific:)

What is your idea of a heavy cpu program?
Are you refering to programs such as F@H or a encoding software?

[reubenb]

ok..
whats ur opinon on the future of Intel? Will AMD succeed it?

[reubenb]

Graphic-Intense... Productionquality editing... etc. etc.

[liorean]

Microsoft will keep Intel alive. Intel will manage to get a 64-bit processor that Microsoft can accept as target platform.

I don't think the future of the Itanium2 is too bright, however. Hammers (Opteron, Athlon64) and IBM PowerPC 970 (G5) will take over it's usage sphere. IA32 will die very slowly, however. I don't see Dell changing to Athlon64 for their main user-end system processor in some time, for example. Especially since none of the 64-bit systems are cold enough for whole-spectrum use yes (laptops don't have the space needed for the heatsinks that you can currently see on 64-bit systems for instance).

However, the company that is the key here is probably ARM. If ARM does something unexpected, such as release a 128-bit RISC with good speed, then Intel may very well licence the core and go along produce something...

If you ask me whether I think AMD will overtake Intel's place as largest PC processor maker, however, I think they very well may.

[reubenb]

yeah..
i see what you mean.
thanks for your help :thumbsup:

[Mhtml]

I was under the impression that 32-bit or less application wouldn't function properly on 64-bit processors. Or is it that they won't fully use the 64-bit capabilities? Either way I see no point in getting a 64-bit processor yet, the market for them is to young in my oppinion and you will be better off waiting till aug or dec next year if you must get one.

[liorean]

As for IA64, yes, you're right about that 32-bit (or more exactly, IA32/X86-32) applications won't work on them. The x86-64 architecture, however, is an extension of the x86-32 architecture, and thus is fully backwards compatible with the 586 and 686 systems. It has the same backwards compatibility problems with 486, 286 and 088 as the 586 and 386 and 286 generations had, respectively.

It is partly true, however, that it won't utilise it's 64-bit capabilities for 32-bit applications. A properly coded modular program, however, should be able to compile one 32-bit and one 64-bit set of base functionality, and chose what to use depending on processor. That makes optimisations possible. Also, this is more a question of compiler optimisations than one of application optimisations. For example, GNU C compiler has optimisations for PPC-32, PPC-64, x86-32 and x86-64 architectures, which means as long as you recompile your programs in 64-bit environment, you'll get the optimisations needed for top performance on that system.


For PC, the 64-bit processors are way too expensive today, I think. What gain you get from them is too small compared to the price difference, especially since they have worse perfomance on 32-bit systems (windows, for instance) and there is virtually no applications that takes advantage of the additional 32 bits.

For Mac, on the other hand, the G5 may very well be worth getting right now.

[oracleguy]

Originally posted by liorean
For PC, the 64-bit processors are way too expensive today, I think. What gain you get from them is too small compared to the price difference, especially since they have worse perfomance on 32-bit systems (windows, for instance) and there is virtually no applications that takes advantage of the additional 32 bits.

I agree, it isn't worth the extra money for one... I'd wait at least a year if not 2, just to see where things end up... it could turn out to be an expensive flop... you just don't know.

Originally posted by liorean
If you ask me whether I think AMD will overtake Intel's place as largest PC processor maker, however, I think they very well may.


I dunno about that... I mean it could happen but not as long as companies like Dell keep selling them.

[reubenb]

yeaa
well theres lots of things not ready for the market like 3g..
is there are 64bit version of windows out yet commercialy available?
there goes another cost factor.
silicon manufacture should consider bundling up software together with the cpu.

;)

[oracleguy]

I think Windows 2000 Dataserver is 64-bit but it also requires at least 8 CPUs or something like that, which obviously isn't an option.

I wouldn't say the market isn't ready for 3Gig CPUs, I mean really you don't need higher than 2gig for most things like email, word, internet, etc.

[liorean]

/me looks at his old 8-bit TI...

[reubenb]

hehe
yeah..
the market isn't really ready for ANYTHING new at this stage...
people still have their old phones for instances without 3G phones that can do everything.

market reviewers need to step up and delay sales for the market's choice.