Sit back and relax, grab your snacks and a glass of coke while I tell you about Intel’s and AMD’s plans of 2007. This is a very long article witch covers very much information, so you might have to get the whole coke bottle to your computer, because this will take some time. But I assure you that it will be worth every second of it, because when you are done reading you will have a lot more knowledge about this than you had when you started reading, and you will know what to expect and what not to expect of the leading chipmakers to PC desktops this year. Well, let’s go ahead and start with the article.
Most of you know that Intel ran past AMD in the race of creating the most powerful processors during the summer of 2006 -- when they released their new microprocessor architecture, Core 2, witch was based on a 65nm fabrication process. A process witch AMD first reached at the end of December, 2006.
Before that summer, AMD had the lead for a period of about 3 years, and had built up a market share of about 85%.
So, the question is, what will happen in this race during 2007? What can we expect from AMD and Intel? Below is the plans of the two competitors, and my conclusion of this whole thing.
AMD
AMD will release a lot of news this year. But many of them are remakes of their old architectures with their new 65nm process, including X2, Sempron, Athlon 64. But they have a hot processor coming for their dual-socket, Socket 1207, witch has four “native” cores. And that combined with a dual-socket means eight cores! Read more about it below.
65nm X2
They have planned to release two new processors with the X2 architecture during the second quarter of 2007, witch will be X2 5200+ and X2 5400+. They will operate at 2,7GHz respectively 2,8GHz.
65nm Sempron
They will also release new single-core Sempron processors fabricated with their new 65nm process. They will introduce the Sempron 3400+, 3500+, 3600+ and 3800+, all during the second quarter of 2007.
65nm Athlon 64
They also have plans on releasing a single core 65nm Athlon 64 processors series, code named “
90nm Quad FX
AMD will continue to produce processors to their Quad FX line, witch based on the dual-socket motherboard 1207. They will be introduced in the second quarter as FX-76, FX-74, FX-72 and FX-70. All of them will be using the old 90nm technique.
"Agena" the monster!
The most interesting processor architecture that AMD will introduce is the “Agena”. It is a performance-oriented 65nm CPU architecture, and is planned to get released in the beginning of the third quarter, 2007. It will be the first quad-core architecture released with all cores on the same die (also called “native” quad-core), either by AMD or Intel.
It will operate at a bus speed of amazing 4.0GHz, witch is double the speed of the previous FX and Athlon 64 processors. It will have a L2 cache of 2MB and run at clock speeds from 2,7 GHz to 2,9 GHz.
The “Agena” processors will likely be introduced under Quad FX-8x model numbers. And since the Quad FX platform, witch the Agena CPU’s will be put on, uses two CPU sockets, AMD will – if nothing goes wrong -- be the first chipmaker in the world to create eight-core processors. At least for PC’s.
Kuma
They will also release a new 65nm native dual-core processor series, mainly for the mainstream market. The current code name for these processors are “Kuma” and it has got a good performance-per-watt ratio. It will run at clock speeds from 2,0 to 2,9GHz and will have 1MB of L2 cache.
Intel
Intel got past AMD in the battle of the best processors last summer, and have slightly increased their marketshare for CPU’s for PC desktops. Will this year be a continued success for intel, or will it be the other way? Its hard to tell, but here’s Intel’s plans for the year.
The new Core 2 Quad series!
In the first week of January, Intel has planned to release their new Core 2 Quad series, witch is Core 2 processors put on a dual-socket motherboard. The models released will be Q6600, Q6400 and Q6300. With clock speeds on 2,4 GHz, 2,13 GHz respectively 1,86 GHz. All of them will have a FSB (Front Side Bus) speed of 1,066 GHz and have a total L2 cache of 8MB (4MB to each of the two processors).
The new Core 2 Due models
In the first half of the year, Intel will release three new processors in their old Core 2 Duo series. The models are E4200, E4300, E4400. With clock speeds on 2,0 GHz, 1,8GHz respectively 1,6 GHz. They will have a FSB of 800 MHz and a shared L2 cache of 2MB. These processors are aimed for the midrange market.
Low price market processors
In the second quarter, Intel will release some processors aimed at the low price market witch is dominated by AMD at the moment. They will be based on the Core 2 architecture, but its unclear if they will make a one core only processor, or do a dual-core with one core disabled. To not confuse costumers the processors will go under the model names Pentium 4 and Celeron, even though they are based on the Core 2 architecture.
The new chip set Bearlake
In the middle of this year, Intel will release a new chip set, code named Bearlake. It will have a FSB speed of 1,333 GHz compared to their previous FSB record speed 1,066 GHz.
At the same time as Intel releases its Bearlake, they will also release a set of new processors, adapted for the Bearlake chip set. They will be based on the Core 2 architecture, and the models released will be 3,0GHz - E6850, 2,66GHz - E6750 and 2,33GHz E6650. The 3GHz mark have only been reached once before by the Core 2 architecture, then it was the Extreme X6800 processor. All of these processors will have a shared L2 cache of 4MB. Another thing that have to be told, to decrease your confusion, is that this Bearlake chip set comes in single-die. That means only one socket.
At the time when Intel releases the Bearlake chip set, and the Bearlake adapted processors, they are also going to release a non Bearlake processor, the E6800 with a speed of 3GHz. The difference between E6800 and the E6850 is the FSB. E6800 has got an FSB of 1,066 GHz and the E6850 has got a FSB of 1,333 GHz.
Intel goes 45 nm!
Well, Intel’s design and technology resources is used for developing a 45nm process. The aim was for finishing it by 2008. But Intel reached their goal faster than anyone would expect. At the end of last year, Intel announced that they successfully had created a prototype of a 45nm processor. They have changed the plans now, and are aiming to release a 45nm processor by the third quarter of this year.
Intel’s fast evolving to the 45nm process will put a big pressure on AMD – who reached the 65nm process at late December last year -- to catch up.
The new architecture that Intel will produce with the 45nm process goes under the code name Penryn. Some rumours say that the Penryn will be based on the old Core 2 architecture, but will take advantage of the new 45nm process to gain larger L2 cache and sharper performance.
The ultimate CPU, Yorkfield!
One of the CPU’s that is planned to be produced with the Penryn architecture is code named “Yorkfield”. Yorkfield will be a native quad-core processor, and has unbelievably 12MB of L2 cache. Clockspeeds of the Yorkfield are not confirmed, but I will release data on them here at the blog as fast as I hear of it.
All of the processors with the Penryn architecture, described above, will be released for the Bearlake chip set (Also described above).
Conclution
This will be a tough battle, but in my eyes Intel has the advantage this year. They will produce their new 45nm processors, if everything goes well – while AMD just discovered the 65nm process. They will have a processor with 12MB L2 cache, and don’t be surprised if they release a dual-socket chip set for that processor. That makes it 24MB L2 cache, and eight cores. This even makes AMD’s monster look like budget. Even though it got eight cores, it only have 2MB of L2 cache. Witch is nothing in comparison with Intel’s 24MB.
So it looks like Intel will be the king of CPU’s this year. And the year after this one AMD might be even more behind, if they don’t figure out how to reach the 45nm process soon.
