Intel says 45nm "Penryn" quad-core chip samples produced

[caveman-jim]

In response to AMD's 45nm process release, Intel says 45nm Quad-core notebook, desktop and server chips will be released sooner:

AMD and IBM have cooperated on chip development since 2003 and last year extended their contract through 2011 to reach the 32nm and 22nm chip generations.

In response, Intel says that it beat AMD to the market with 65nm chip sales by more than a year and has already begun producing samples of its 45nm "Penryn" quad-core chip for notebooks, desktops and servers. Intel plans to ship those chips in the second half of 2007.

Even if AMD improves its manufacturing, Intel is poised to continue its lead in shrinking chip geometries because it has shorter gate lengths and SRAM (static RAM) cell sizes, said Rob Willoner, a technology analyst with Intel's technology and manufacturing group. By shrinking from 90nm to 65nm, Intel chips have double the transistor density while reducing leakage fivefold (at constant performance) and requiring 30% less switching power.

"People have different meanings when they talk about the 65nm process. There's not an industry standard," he said. AMD's Opteron chip was a commercial success because its excellent design compensated for poor dimensions at the 90nm scale, he said.

Source: ComputerWorld

[Targon]

Intel may hold a lead when it comes to process technology, but one thing that will be interesting to see is how things turn out when it comes to total system design and performance.

At this point in time, Intel has a huge advantage in terms of processor performance. That advantage is enough to give Intel the performance advantage, even though their overall system design isn't as sophisticated. When K8L is finally released though, Intel's CPU design may not give it the lead in terms of overall CPU performance. At that point, we will get to see which company will have the best system design.

Does it really matter if Intel is at 45nm before AMD if AMD based machines are faster in many/most tasks? AMD has proven that when it comes to innovation, they tend to be ahead of Intel in terms of system design. AMD will easily be in the lead when it comes to integrating the GPU into the system for improved performance in another two years.

So, we will need to see if the 45nm Intel processors will be enough to keep Intel ahead of AMD when you look at the overall system performance. We saw this in the Pentium 4 era where Intel was able to push the Pentium 4 very well, but when it came to overall system performance, AMD had the advantage. We may see this case repeat itself where Intel will push their 45nm processors, but in overall system performance, AMD will take the crown.

[gamefoo21]

Originally Posted by Targon

Intel may hold a lead when it comes to process technology, but one thing that will be interesting to see is how things turn out when it comes to total system design and performance.

At this point in time, Intel has a huge advantage in terms of processor performance. That advantage is enough to give Intel the performance advantage, even though their overall system design isn't as sophisticated. When K8L is finally released though, Intel's CPU design may not give it the lead in terms of overall CPU performance. At that point, we will get to see which company will have the best system design.

Does it really matter if Intel is at 45nm before AMD if AMD based machines are faster in many/most tasks? AMD has proven that when it comes to innovation, they tend to be ahead of Intel in terms of system design. AMD will easily be in the lead when it comes to integrating the GPU into the system for improved performance in another two years.

So, we will need to see if the 45nm Intel processors will be enough to keep Intel ahead of AMD when you look at the overall system performance. We saw this in the Pentium 4 era where Intel was able to push the Pentium 4 very well, but when it came to overall system performance, AMD had the advantage. We may see this case repeat itself where Intel will push their 45nm processors, but in overall system performance, AMD will take the crown.

So says the amd fanboy...

You do realize that its always a back and forth war in terms of innovation. You can't say that AMD has always been more foreward thinking. K5 was a pentium knock off, k6 was another pentium knock off. I had a celeron 366 that would trounce the a 450 k6-2.

K7 was a design they bought. K8 was a K7 but with more technology they bought and bolted on. K8 is really a tweaked K7 with an integrated memory controller. That memory controller was a total dickwad to deal with until the venice and similar gen cores and still isn't as flexible as the intel solution.

Need I mention the whole L3 and cross bar for K8L? Intel has dynamic cache sharing and straight through communication through the L2. AMD can't buy it so they can't beat it.

AMD isn't an innovator by any means, what they do is they take something thats already there put a spin on it and say its there own. Hell they can't even build a cpu if it wasn't for Ibm bailing them out.

Now... If you want to get on my case about the Pentium 4, it was rushed was never meant to be in .18um format(Willamette), the B series pulled it into a tie with the Athlon XP, the C series swatted the Bartons like pesky flies. So AMD went into uber cheap mode. Then the athlon 64 came out, and AMD held the desktop crown.

Then Intel put out Banias, and later the Dothan and thoroughly stomped on AMD in the laptop market, and even in the desktops for those willing to invest in a converter so they could strap em a dothan and run it with crazy clocks.

[AllexxisF1]

So says the Intel Fanboy

How on Earth can you say Intel is the innovation leader when they are STILL using an FSB architecture and UNGODLY amounts of Cache .

Targon was absolutely spot on, Intel might make it to market first but make no mistake AMD is at the forefront of processor technology.

[Grim Reaper,The]

Originally Posted by Targon

Intel may hold a lead when it comes to process technology, but one thing that will be interesting to see is how things turn out when it comes to total system design and performance.

At this point in time, Intel has a huge advantage in terms of processor performance. That advantage is enough to give Intel the performance advantage, even though their overall system design isn't as sophisticated. When K8L is finally released though, Intel's CPU design may not give it the lead in terms of overall CPU performance. At that point, we will get to see which company will have the best system design.

Does it really matter if Intel is at 45nm before AMD if AMD based machines are faster in many/most tasks? AMD has proven that when it comes to innovation, they tend to be ahead of Intel in terms of system design. AMD will easily be in the lead when it comes to integrating the GPU into the system for improved performance in another two years.

So, we will need to see if the 45nm Intel processors will be enough to keep Intel ahead of AMD when you look at the overall system performance. We saw this in the Pentium 4 era where Intel was able to push the Pentium 4 very well, but when it came to overall system performance, AMD had the advantage. We may see this case repeat itself where Intel will push their 45nm processors, but in overall system performance, AMD will take the crown.

What a joke, in two years amd might exist only in some poor sod's sobbing memories while Intel is already sampling 45nm based CPUs right now.
Besides you're nuts if you think that Intel will still be on 45nm in two years, especially considering how aggressive they have become.

Originally Posted by AllexxisF1

...make no mistake AMD is at the forefront of processor technology.

Sure they are, just look how they are owning the competition...oh, sorry my mistake.

[AllexxisF1]

Originally Posted by Grim Reaper,The

What a joke, in two years amd might exist only in some poor sod's sobbing memories while Intel is already sampling 45nm based CPUs right now.
Besides you're nuts if you think that Intel will still be on 45nm in two years, especially considering how aggressive they have become.

You think AMD will no longer exist in two years

Sure they are, just look how they are owning the competition...oh, sorry my mistake.

Owning the competition with it's THIRD answer to AMD's K8. All the while without an onboard memory controller, and the ungodly amounts of onboard cache.

I'll be sure to save your post to when the benches come out for K8L Mr. Fancy Pants.

[Grim Reaper,The]

Originally Posted by AllexxisF1

You think AMD will no longer exist in two years :nuts:

I didn't say they wont exist, I said that they MIGHT not exist. If your fanboyism makes you too dense to understand the difference then you have a problem.
Few companies can survive mistakes in a multi billion market and AMD is hardly one of them and there is no guarantee that they wont make any.

Owning the competition with it's THIRD answer to AMD's K8. All the while without an onboard memory controller, and the ungodly amounts of onboard cache.

I'll be sure to save your post to when the benches come out for K8L Mr. Fancy Pants.

P4 was Intel's expensive mistake, sure, however, they have survived it.

Even without an on die memory controller C2D is vastly superior to anything AMD has to offer with an on die memory controller. And nothing stops amd from throwing "ungodly" amounts of cache on their CPUs, unless, of course, they know it will do them no good.

As far as I'm concerned you can etch that post into you forehead, I don't care either way.

[LuCi-_-]

Originally Posted by gamefoo21

So says the amd fanboy...

You do realize that its always a back and forth war in terms of innovation. You can't say that AMD has always been more foreward thinking. K5 was a pentium knock off, k6 was another pentium knock off. I had a celeron 366 that would trounce the a 450 k6-2.

K7 was a design they bought. K8 was a K7 but with more technology they bought and bolted on. K8 is really a tweaked K7 with an integrated memory controller. That memory controller was a total dickwad to deal with until the venice and similar gen cores and still isn't as flexible as the intel solution.

Need I mention the whole L3 and cross bar for K8L? Intel has dynamic cache sharing and straight through communication through the L2. AMD can't buy it so they can't beat it.

AMD isn't an innovator by any means, what they do is they take something thats already there put a spin on it and say its there own. Hell they can't even build a cpu if it wasn't for Ibm bailing them out.

Now... If you want to get on my case about the Pentium 4, it was rushed was never meant to be in .18um format(Willamette), the B series pulled it into a tie with the Athlon XP, the C series swatted the Bartons like pesky flies. So AMD went into uber cheap mode. Then the athlon 64 came out, and AMD held the desktop crown.

Then Intel put out Banias, and later the Dothan and thoroughly stomped on AMD in the laptop market, and even in the desktops for those willing to invest in a converter so they could strap em a dothan and run it with crazy clocks.

Don't get me wrong, I like the new core 2 duo's because of their overclockability and great performance. But don't for a second act like it was forward thinking. Dothan? The dothan was a tweaked p3 core. The core 2 duo, is still a highly modified/tweaked p3 core. Get off the high horse. Intel knew their p4 netburst architecture sucked balls, thats why they went back to their more efficient pipeline of the p3. No p4 processor ever spanked an amd 64 aside from a few tests that favored long pipelines like encoding/decoding.

Core 2 duo is the first processor since the p3 that intel has done right, their netburst architecture was an absolute failure and intel was only slightly able to cover it up with clock speed boosts and heat problems were all over the place when they pushed around 4.0ghz.

All that said, I still would buy a core 2 duo over anything else that exists today. I'm of the boat , at least in processors, that goes for the best at the moment and intel can hold that flag for a while.


P6 based chips

• Celeron (Pentium II/III Derivative)
• Celeron M (Banias/Dothan Derivative)
• Pentium Pro
• Pentium II
• Pentium II Xeon
• Pentium III
• Pentium III Xeon
• Pentium M
• Intel Core

[sandorski]

Intel has always been ahead when it comes to shrinking the process. It certainly is an advantage, but has never been enough of an advantage to drastically change things in regards to Intel vs AMD. There are far more important things.

[Targon]

I think you missed half of my post, because you are STILL looking at processor, processor, processor, when my post looked beyond JUST the processor. I also didn't say anything about AMD catching up at the 45nm process level, or even the next point beyond that.

Let me try this again, and make this very simple for you so you don't miss it.

Point 1: Intel currently holds the lead in single core performance, and this extends to dual and quad core performance. I did not deny this fact in any way.

Point 2: Intel has not done a significant change to the overall system architecture in quite some time. Intel may have improved things, but the overall system design is still very similar to what we have seen for the past 20 years, just with updates. New slot types, faster speeds, etc. A point to point bus design like HyperTransport has been avoided by Intel because it would be an obvious case of Intel following an AMD lead.

Point 3: AMD KNOWS that they can't keep up with Intel when it comes to R&D dollars, so the people at AMD MUST look beyond process technology in order to compete.

Point 4: IF(and that's a big if) AMD releases a processor that matches the performance of the processors Intel has out, the system architecture differences that AMD has implemented would make an AMD system perform better. Note that because Intel has a huge CPU performance advantage right now, AMD is behind in overall system performance. This will probably change when K8L is released.


So, from this, while I may be an AMD fan, I am far from being unrealistic about the market. At no time have I said that AMD would suddenly re-take the performance lead with K8L, though I have implied that I feel it's a good possibility. Only time will tell, but Intel needs to understand that there is more to system performance than just the CPU.

CPU innovation != system design innovation.

Originally Posted by gamefoo21

So says the amd fanboy...

You do realize that its always a back and forth war in terms of innovation. You can't say that AMD has always been more foreward thinking. K5 was a pentium knock off, k6 was another pentium knock off. I had a celeron 366 that would trounce the a 450 k6-2.

K7 was a design they bought. K8 was a K7 but with more technology they bought and bolted on. K8 is really a tweaked K7 with an integrated memory controller. That memory controller was a total dickwad to deal with until the venice and similar gen cores and still isn't as flexible as the intel solution.

Need I mention the whole L3 and cross bar for K8L? Intel has dynamic cache sharing and straight through communication through the L2. AMD can't buy it so they can't beat it.

AMD isn't an innovator by any means, what they do is they take something thats already there put a spin on it and say its there own. Hell they can't even build a cpu if it wasn't for Ibm bailing them out.

Now... If you want to get on my case about the Pentium 4, it was rushed was never meant to be in .18um format(Willamette), the B series pulled it into a tie with the Athlon XP, the C series swatted the Bartons like pesky flies. So AMD went into uber cheap mode. Then the athlon 64 came out, and AMD held the desktop crown.

Then Intel put out Banias, and later the Dothan and thoroughly stomped on AMD in the laptop market, and even in the desktops for those willing to invest in a converter so they could strap em a dothan and run it with crazy clocks.

[Targon]

AMD is up there in terms of performance, but as I said in my other recent post, AMD knows when they can't compete in terms of things like clock speed, so they look for other ways to improve performance.

I do feel that Intel NEEDS huge amounts of cache to get decent performance out of their processors, and that will hurt them in the long run. Intel has tended to release dual core parts that were two single core processors kludged together to respond to true dual core parts from AMD, and to release a kludge of a quad core chip to beat AMD to market. When it comes right down to it though, it takes Intel quite a bit longer to release a "proper" response to AMD leadership in terms of innovation.

AMD beat Intel with a true dual-core design. AMD could have released K8L on a 90nm process, but decided not to because the power demands would tarnish the improvements to the design. Do you honestly think that an Intel processor with only 512KB/core would be able to hold the lead going forward? How about 1MB/core? Through the entire P4 era, huge amounts of cache were required, and even with the Core design, you see a huge amount of cache being required.

So, what's next from Intel, one gigabyte of cache per core being required to keep Intel ahead because AMD has managed to stay in the race?

System architecture needs to be looked at. Intel can have their hundreds of cores, it won't make any one application run any faster.

Originally Posted by AllexxisF1

So says the Intel Fanboy

How on Earth can you say Intel is the innovation leader when they are STILL using an FSB architecture and UNGODLY amounts of Cache .

Targon was absolutely spot on, Intel might make it to market first but make no mistake AMD is at the forefront of processor technology.

[Targon]

Originally Posted by Grim Reaper,The

What a joke, in two years amd might exist only in some poor sod's sobbing memories while Intel is already sampling 45nm based CPUs right now.
Besides you're nuts if you think that Intel will still be on 45nm in two years, especially considering how aggressive they have become.


Sure they are, just look how they are owning the competition...oh, sorry my mistake.

In two years, AMD will be on 45nm, Intel will be on to something else. But if AMD has a way to accelerate graphics performance, an AMD system may SEEM like it is faster. Intel hasn't indicated that they have anything really innovative when it comes to system design, only CPU design.

In two years, we may see Intel has run into problems with their new process technology, or two years after that. It's fine for a company to say they plan to improve this or that, but do you remember how Intel was talking about taking the P4 to 5GHz?

[candle_86]

Originally Posted by gamefoo21

So says the amd fanboy...

You do realize that its always a back and forth war in terms of innovation. You can't say that AMD has always been more foreward thinking. K5 was a pentium knock off, k6 was another pentium knock off. I had a celeron 366 that would trounce the a 450 k6-2.

K7 was a design they bought. K8 was a K7 but with more technology they bought and bolted on. K8 is really a tweaked K7 with an integrated memory controller. That memory controller was a total dickwad to deal with until the venice and similar gen cores and still isn't as flexible as the intel solution.

Need I mention the whole L3 and cross bar for K8L? Intel has dynamic cache sharing and straight through communication through the L2. AMD can't buy it so they can't beat it.

AMD isn't an innovator by any means, what they do is they take something thats already there put a spin on it and say its there own. Hell they can't even build a cpu if it wasn't for Ibm bailing them out.

Now... If you want to get on my case about the Pentium 4, it was rushed was never meant to be in .18um format(Willamette), the B series pulled it into a tie with the Athlon XP, the C series swatted the Bartons like pesky flies. So AMD went into uber cheap mode. Then the athlon 64 came out, and AMD held the desktop crown.

Then Intel put out Banias, and later the Dothan and thoroughly stomped on AMD in the laptop market, and even in the desktops for those willing to invest in a converter so they could strap em a dothan and run it with crazy clocks.

just to adjust your history of events, K5 was a Pentium Class chip, but it wasnt a knock off, AMD tried to build one from scratch and it didnt work. The K6 was the inhouse project of a company AMD bought out, i dont remember the name now, also in bussiness apps the K6 was faster than similar clocked Pentiums, but lost in Floating point calculations, total opposite of today. The K7 was an in house AMD project, they didnt buy the core, the devloped it them selves, and in Interger and Floating Point it creamed the Pentium III, untill Coppermine launched and AMD countered with full speed L2 on there Thunderbird line of chips and agian was winning. The K7 and K8 are similar indeed, but then agian the Core2 is a heavly tweaked Pentium Pro, whats your point, why retire something thats good? Now your history is fixed

[Grim Reaper,The]

Originally Posted by Targon

Point 2: Intel has not done a significant change to the overall system architecture in quite some time. Intel may have improved things, but the overall system design is still very similar to what we have seen for the past 20 years, just with updates. New slot types, faster speeds, etc. A point to point bus design like HyperTransport has been avoided by Intel because it would be an obvious case of Intel following an AMD lead.

And HyperTransport is an evolutionary step to the FSB same as PCIe to PCI, no more than an update to a concept.

Intel works on its own FSB update, it's called "Common System Interface", which should benefit Intel far more than HyperTransport since CSI will allow them, beside all the expected improvements like speed etc., to use Xeons and Itaniums on the same motherboard.

Also, why do you assume that slapping a GPU next to the CPU will give AMD any serious advantage ?
GPU is pass through device, it gets the data, processes it and outputs to the screen. And since it gets most of that data from it's own RAM having a fast connection to the CPU will give marginal improvements at best.