Core after Prescott

[tgxiii]

Does anyone know the name of the Intel core that is coming out in 2004, the one that will come out after the Prescott? I don't know if it was a rumor, a hoax, or if it was true but I read something about it somewhere. Also, if you know websites that have information on them please post them. Thank you.

[beppodmime]

According to www.theinquirer.net , Intel's next desktop chip will be code-named Tejas. It is supposed to use the 774 pin LGA packaging, and will start at speeds of around 4.4 GHz. It will feature new instructions, will be a 90 nanometer part, and is slated to be released at the end of next year. We'll see how that goes.

Rumors abound on Intel working on a "secret" 64 bit desktop/workstation processor code-named Yamhill. It's all speculation and rumor, as Intel has repeatedly said in the past that they will never produce a 64 bit chip that's backwards compatible with 32 bit extensions. People's take on this generally leans towards this chip being Intel's backup plan in the event that the AMD-64 architecture ends up taking control of the market.

Other than that, they are working on the Itanium 2 enterprise processor, a 64 bit server chip with a totally new instruction set. These chips are based on the IA-64 instruction set, and will be unable to run current X86 code.

For rumors and such, theinquirer.net is pretty much the only source I check, but I don't really play into rumors too much, so I'm not in the habit of checking. I check www.hardocp.com regularly for news and events, and every once in awhile they'll link to interesting articles from the Inquirer.

[tgxiii]

Wow, thanks for all the info.

[Telexen]

Originally posted by beppodmime

Rumors abound on Intel working on a "secret" 64 bit desktop/workstation processor code-named Yamhill. It's all speculation and rumor, as Intel has repeatedly said in the past that they will never produce a 64 bit chip that's backwards compatible with 32 bit extensions. People's take on this generally leans towards this chip being Intel's backup plan in the event that the AMD-64 architecture ends up taking control of the market.

The Prescott has been in development nearly as long as the Athlon64/Opteron...and

[Xtreeme]

Thats sad. As the opteron is a hole new ball game. Completely redesigned. The Prescott only added a couple messily instruction sets and aimed for 90nm. Other than that its the same old P4 with higher mhz. Oh and they added 2 mor stages to 20stage pipe cause of the heat. That will slow it even more per mhz! LOL. And it took them just as long to do THAT> why the 100watt die then. I mean with all that time the power prob. should have been #1 concern.

[beppodmime]

Well, the Opteron isn't really all that architecturally different from the latest T-bred/Barton. All they did to that core was add SSE2 instructions, add the 64-bit registers, and add cache. Well, that and integrate the memory controller into the processor itself. They really improved the processor where most believe it needed to be improved, though, and left the good stuff alone. The P4 is able to run at such high MHz mostly due to the length of it's pipe. True that on the power/heat concerns, though. Them's some hot potatoes, although I'm sure Intel will be working that out soon enough.

Intel is working on the power/heat issues. The latest internal roadmap that I have lists both the mobile Prescott-M 3.73 GHz model and desktop Prescott-256K 3.06 GHz around the 80W range, a serious cut from the 100+W for the first round of Prescott engineering samples.

Robert Richmond

[Xtreeme]

Oh-that is a good power cut. Least they are handling it the right way.

"Well, the Opteron isn't really all that architecturally different from the latest T-bred/Barton."

hmm define architecture! 64bit enhanced-32bit backward compat. AND uses integrated NB into cpu die. That is a completely NEW architecture with the old instruction reused as always (they all do this for compat with old progs and each other too). Only non architectual change was adding SSE2 which is just a instruction set embeded into cpu code. This is FAR from Intel's move to just a faster, smaller core.

"All they did to that core was add SSE2 instructions, add the 64-bit registers, and add cache. Well, that and integrate the memory controller into the processor itself."

Hmm added sse2 is just little thing I agree. But integrating mem controler is revolutionary stuff-how can you think that didnt take TONS of time. THats a HUGE thing. I dont see Intel integrating their NB into the cpu die. Of course not that would cut into their chipset market segment. LOL. Adding 64bit to the cpu is more than you make it sound too. THis is also revolutionary tech. I mean Intel still is going only full 64 or 32bit on cpu's . None are doing both like opteron-Intel wont for another 2 YEARs. THats from a Intel official representive's statement on the .net.

All in all. I think AMD did alot more work getting opteron out the door then Intel did with only having to deal with a 100watt power issue.

Having said all that. I will say this. I think AMD has tons more resting on the success of the opteron. Without it we may loose amd. So they HAD to get on the ball or loose big time. Intel once again isnt being rushed and is showing it by taking their time with improvments. Remember without AMD right there-Intel was always taking her time going up in speed and enhancments. Back when AMD was not as well known-we payed ney on $500 for a 30mhz boost!