Intel enters billion-transistor processor era

[sonoma1993]

I was just reading an article over on amdzone.com website, that intel is plaaning a goal for 2005 to release a 1 billion transistor Itanium 2. Also some anounce that amd is doomed because intel has reached the key 1 billion transistor mark

amd zone- intel 1 billion

ee times- intel 1 billion

 

[coldpower27]

I serious doubt AMD is doomed because Intel released a 1 Billion transistor processor.

They are talking about the Dual Core Hyperthreading enabled Montencito Core, 1.72 Billion Transistors on 90nm process, 2x1MB of LV2, 2x12MB LV3 for over 26MB of total cache. Very impressive for 90nm technology.

 

[sonoma1993]

yeah that what im thinking coldpower,

 

[evilharp]

Originally posted by: sonoma1993
Also some anounce that amd is doomed because intel has reached the key 1 billion transistor mark

Doomed?

Lets see, Intel samples a 1.27 billion transistor chip that nobody wants... Most of the transistors are a result of cache... Intel desktop chips double as hotplates...

I think not. Intel's super chip, the Itanium 2, is now in the dreaded "big iron" category. Big Iron = Small Sales (at a huge markup). No threat in the desktop/laptop to mid-range server space (AMD's turf).

 

[TuxDave]

Originally posted by: evilharp

Originally posted by: sonoma1993
Also some anounce that amd is doomed because intel has reached the key 1 billion transistor mark

Doomed?

Lets see, Intel samples a 1.27 billion transistor chip that nobody wants... Most of the transistors are a result of cache... Intel desktop chips double as hotplates...

I think not. Intel's super chip, the Itanium 2, is now in the dreaded "big iron" category. Big Iron = Small Sales (at a huge markup). No threat in the desktop/laptop to mid-range server space (AMD's turf).

Comments like this is what I think bothers me the most. I dread the day that when the processor I'm working on is released that I'll get to enjoy all the negative comments from this forum.

 

[Lonyo]

Originally posted by: TuxDave

Originally posted by: evilharp

Originally posted by: sonoma1993
Also some anounce that amd is doomed because intel has reached the key 1 billion transistor mark

Doomed?

Lets see, Intel samples a 1.27 billion transistor chip that nobody wants... Most of the transistors are a result of cache... Intel desktop chips double as hotplates...

I think not. Intel's super chip, the Itanium 2, is now in the dreaded "big iron" category. Big Iron = Small Sales (at a huge markup). No threat in the desktop/laptop to mid-range server space (AMD's turf).

Comments like this is what I think bothers me the most. I dread the day that when the processor I'm working on is released that I'll get to enjoy all the negative comments from this forum.

Itanium is a totally different processor to most things, and isn't hugely popular or important. IIRC it's also been losing ground to AMD64 (Opterons).

 

[TuxDave]

Originally posted by: Lonyo
Itanium is a totally different processor to most things, and isn't hugely popular or important. IIRC it's also been losing ground to AMD64 (Opterons).

Yeah, I'm pretty familiar with Itanium and its less that luster success story. In fact, I know for sure there is one AT member who worked on it too.

 

[Maluno]

...who??? are you an engineer?

 

[GEOrifle]

for INTEL gona be better if he spends money to introduce some kind REFREGERATOR for
it's damn PRESCOT cpu, i have brand new 478 MOB and i don't know what the hell am suppost to do--trow in the garbege i guess, and switch to AMD.

 

[jiffylube1024]

Originally posted by: sonoma1993
I was just reading an article over on amdzone.com website, that intel is plaaning a goal for 2005 to release a 1 billion transistor Itanium 2. Also some anounce that amd is doomed because intel has reached the key 1 billion transistor mark.

Really? That's the stupidest thing I've ever heard! a 1B transistor core shows just how behind Intel really is. They've had to over-cache the P4's to even come close to competing in the desktop arena, and Itanium (always bloated) is even more overloaded with transistors.

I certainly attest that this is impressive, but can you imagine the yields on this thing? Yikes. Intel certainly is the king in terms of being able to pull off the impossible (taking Prescott to near 4 Ghz, putting more cache than thought economically feasible into all of their chips, etc), but this goes to show what a tough time AMD really is giving them.

Although truth be told, Itanium is a super-niche chip. It's already ridiculously expensive, and this will probably drive costs up even more.

This is akin to one of those "eating 10 burgers in 1 sitting" milestones. It's an impressive achievement, but I wouldn't go bragging about it to everyone...

 

[zephyrprime]

Originally posted by: jiffylube1024Really? That's the stupidest thing I've ever heard! a 1B transistor core shows just how behind Intel really is. They've had to over-cache the P4's to even come close to competing in the desktop arena, and Itanium (always bloated) is even more overloaded with transistors.

I certainly attest that this is impressive, but can you imagine the yields on this thing? Yikes. Intel certainly is the king in terms of being able to pull off the impossible (taking Prescott to near 4 Ghz, putting more cache than thought economically feasible into all of their chips, etc), but this goes to show what a tough time AMD really is giving them.

Although truth be told, Itanium is a super-niche chip. It's already ridiculously expensive, and this will probably drive costs up even more.

This is akin to one of those "eating 10 burgers in 1 sitting" milestones. It's an impressive achievement, but I wouldn't go bragging about it to everyone...

Yeah, Itantium has good performance in number crunching apps but it's really pretty unimpressive when you consider the humongeous transistor cost required to make the itanium.

 

[Accord99]

Originally posted by: jiffylube1024

I certainly attest that this is impressive, but can you imagine the yields on this thing? Yikes. Intel certainly is the king in terms of being able to pull off the impossible (taking Prescott to near 4 Ghz, putting more cache than thought economically feasible into all of their chips, etc), but this goes to show what a tough time AMD really is giving them.

Montecito is not targetted at Opteron, it's targeted at the Power5 series which features on-die memory controllers and extremely large caches of its own. The big caches are important for scalability for the large way systems that they're used in.

 

[DrMrLordX]

It's not how many transistors you have, but what you do with them, that counts.

 

[dmens]

Originally posted by: TuxDave
Comments like this is what I think bothers me the most. I dread the day that when the processor I'm working on is released that I'll get to enjoy all the negative comments from this forum.

Why? That comment was moronic to begin with. All the cpu designers I work with don't get paid enough to care about the outcome of the cpu itself, never mind some know-nothing fanboi on a web foum, lol.

 

[TuxDave]

Originally posted by: dmens

Originally posted by: TuxDave
Comments like this is what I think bothers me the most. I dread the day that when the processor I'm working on is released that I'll get to enjoy all the negative comments from this forum.

Why? That comment was moronic to begin with. All the cpu designers I work with don't get paid enough to care about the outcome of the cpu itself, never mind some know-nothing fanboi on a web foum, lol.

lmao... true enough.

 

[IntelUser2000]

Yield shouldn't be a problem since most of the 596mm2 die size is cache and they are protected from failures by mechanisms to correct errors like ECC and ability to turn off defective portions. Its actually 1.72 billion transistors, 57 million transistors which belongs to two cores. Despite so many transistors and at 90nm, it will be 100W at 2.0GHz.