Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Originally posted by SiliconAddict
Where's AMD in all of this?!?! :confused:

My thoughts exactly. Though how long they stay with 130nm depends on how hight they can push the current generations of 64 prcessors (Opterons and Athlon 64.) A 90nm Opteron would be something to smile about:)

Remember that the Athlon 64FX (2.2GHz) is only marginally slower than a P43.2EE in 32-bit mode, wait till both the G5 and AMD chips are running wide open in 64-bit mode with oodles of RAM.
 
Originally posted by manitoubalck
....
Remember that the Athlon 64FX (2.2GHz) is only marginally slower than a P43.2EE in 32-bit mode, wait till both the G5 and AMD chips are running wide open in 64-bit mode with oodles of RAM.

Would that really boost up normal perforance? My understaning is that the intel is simply a faster chip. 3.2 intel with 4 gigs should run cirlces around the 2.2 amd with 8 gigs. Maybe I am mistaken, but i dont think the relative lack of ram is causing THAT much of a bottleneck for the current intel.
 
Intel AMD

It would seem to me that AMD would have a really tough time winning in the chip war. The specs for their processors are a little low compared to Intel in numbers, if not performance, and the majority of the serious peecee people I know see AMD procs as inferior (kinda like they do to Apple...) I would bet that Intel keeps dominating the PC market, with (hopefully!) some more serious competition from Macs using IBM.

Of course, this is all speculation, especially considering how often chip-makers don't come through on time or at all just consider that we were supposed to see Prescott in late August to September, then October, now Q1 2004... and this .65 is 2005- seems like just marketing hype to me
 
Re: Intel AMD

Originally posted by IIvan
It would seem to me that AMD would have a really tough time winning in the chip war. The specs for their processors are a little low compared to Intel in numbers, if not performance, and the majority of the serious peecee people I know see AMD procs as inferior (kinda like they do to Apple...)

Eh? AMD chips are the darlings of the high performance computing crowd, and quite popular with gamers as well. Clock-for-clock, they school Intel chips. There were some problems with motherboard reliability early on though. AMD floating point performance has been superior to Intel's for some time now, which has made them popular for supercomputing clusters.

For most intents and purposes, a 2.0 GHz Opteron wipes the floor with Intel processors that are clocked much higher. AMD chips not only some of the fastest you can buy per clock, they are also among the very fastest you can buy in absolute terms as well. Intel chips are only fast in narrow domains, and even then at insane clock rates. The G5 is generally a bit slower than an Opteron clock-for-clock, but it will make a better general purpose processor than Intel's current lineup and has a much better architecture than the P4 in its various incarnations.
 
Re: Re: Don't write off Moto yet

Originally posted by dongmin
Can't believe people are referencing old Moto roadmaps, especially one cooked up by the Register.

The older Register story was written back before SNDF 2003 which is where Motorola released info on the 7447/57 and their future plans, including dual core.

Unfortunately, the relevant presentation is now not available. H1112 by Chuck Corley although the first one does mention dual cores.

http://e-www.motorola.com/webapp/sps/site/overview.jsp?nodeId=02VS0llCc5pzMPYZjglff5tfP8n1kKD9fk
 
Re: Re: Intel AMD

Originally posted by tortoise
...The G5 is generally a bit slower than an Opteron clock-for-clock...

Not sure this is true in all domains-- judging by the Top 500 list we've all been following 2200 2GHz G5s had a faster Rmax and Rpeak than 2800+ 2GHz Opterons...

Yeah these are cluster benchmarks, but I think Rpeak is a theoretical number ignoring interconnect issues. This also neglects SIMD performance (I don't know how that stacks up).

Of course this is all from memory and I'm too lazy to find the source so I might be wrong...
 
Originally posted by dho
Would that really boost up normal perforance? My understaning is that the intel is simply a faster chip. 3.2 intel with 4 gigs should run cirlces around the 2.2 amd with 8 gigs. Maybe I am mistaken, but i dont think the relative lack of ram is causing THAT much of a bottleneck for the current intel.

I am very, very curious as to from where you recieved that impression? The general consensus right now is that the AMD Athlon 64 FX-51 has the edge in gaming and similar applications, while the P4 EE does media encoding, etc. better.

Regards, GulGnu

-Stabil som fan!
 
Re: Intel at 65-nm in 2005?

Originally posted by Macrumors
eeTimes and ZDNet are reporting on Intel's push into the 65-nm chip design space.

Intel announced that they are targeting 65-nm chips in 2005 with SRAM samples at this time. ZDNet provides a good overview of the advantages and challenges in reducing chip size. In short, a smaller chip "improves performance, reduces costs and can potentially cut energy consumption."

Current chips, including the PowerPC 970, have been produced at 130-nm, but both Intel and IBM are ramping up 90-nm chip production at this time. 90nm PowerPC chip technology will be presented in February 2004 by IBM and will presumably make their way into future Macintoshes.

As previously reported IBM also plans on introducing 65-nm chips in 2005 using their SSOI (Strained Silicon on Insulator) technology.

whew, about time

i don't want to hear about "heat" issues anymore and going to .065 microns will be great for the laptops especially and i assume it will improve battery time

as it stands now, even the best laptops have to be recharged nightly by a student if they reasonably use it for school and gaming, which i know will happen:p
 
Re: Re: Intel AMD

Originally posted by tortoise
... Clock-for-clock, they school Intel chips.
True, as is the case with the G5 - still, this only really matters if you're competing clock-for-clock rather than chip-for-chip. IBM and AMD get more done per cycle, intel traditionally offers higher numbers of cycles.
AMD floating point performance has been superior to Intel's for some time now, which has made them popular for supercomputing clusters.
Again a similar situation to the IBM chips. Unfortunately, for an awful lot of applications its integer performance that really counts, which is why intel's poor specFP showings have never really hurt them outside the research and supercomputer setting. Moving to 64-bit will make this even more critical, as some applications that currently use floating-point will be able to move to integer-based designs, which are generally more accurate.
For most intents and purposes, a 2.0 GHz Opteron wipes the floor with Intel processors that are clocked much higher. AMD chips not only some of the fastest you can buy per clock, they are also among the very fastest you can buy in absolute terms as well.
That's a better comparison. Comparing power per clock is like looking at two engines are comparing Horsepower-per-liter -- of great interest to engine builders, pretty much irrelevant to most people/companies.
Intel chips are only fast in narrow domains, and even then at insane clock rates.
Well... again, nobody really cares about the clock rates being high. And the "narrow domains" are those that matter the most to a lot of consumers and businesses (integer processing).
The G5 is generally a bit slower than an Opteron clock-for-clock, but it will make a better general purpose processor than Intel's current lineup and has a much better architecture than the P4 in its various incarnations.
I'm interested in seeing how the PM's replacement does - these are built from the old PIII lineage, remember, and the current sets in Centrino laptops are remarkably powerful, seemingly performing clock-for-clock with the G5 in the tests that I've seen and done. I don't know what kind of yields they're getting of these chips, but can't help but wonder what a high-power version would be like. Still, right now intel can churn out very fast P4s for very little money, which is probably the best solution for most consumers. Even businesses, once you get out of the rendering type environments are better served with their historically strong integer performance. This is the area where I'd really like to see IBM and AMD make some strides.

-Richard
 
Bring it on Mutha Fugga!

Big ****ing deal. Didn't Intel say they were going to be in the 64 bit area for desktops too?

Big things generally move slow in dramatic change. Like Microsoft.
IBM can change faster than Intel can. Even if they get to 65, I'm sure IBM could get there before then. IBM has prototypes for products we won't see for two years. I'm sure we'll be there.

Maybe by then more people will unplug themselves from the Bull**** world of Wintel.

This might get interesting. Know what this means IBM? WORK FASTER DAMNIT! :D
I want my ****ing G6 by Winter 04!

Which is faster, G5 or AMD?
I didn't get the MacWorld yet.
 
Originally posted by Drifter
Bring it on Mutha Fugga!

Big ****ing deal. Didn't Intel say they were going to be in the 64 bit area for desktops too?

...

Which is faster, G5 or AMD?
I didn't get the MacWorld yet.

i think overall the PC world still has the speed crown according to that issue

but not as by much when we only had the G4 on our side

if ibm and apple keep on track, or actually, better than expected, then hopefully by late next year or in 2005 we will catch up with the wintel/win-amd world

the next challenge after that will be to get prices down and capture more market share...he he...like ten percent since our long standing five percent of the market has mostly been under four percent in recent years...blah
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.