Intel Kills Plans For 4GHz Pentium

[KingFish]

Intel is dumping plans to release a Pentium 4 processor that runs at 4GHz, saying it will boost performance on next year's chips using other means than clock speed.

The Santa Clara, Calif.-based chip company said it plans to brief PC manufacturers Thursday on the latest changes to its processor road map. The main change is that the 4GHz Pentium 4--scheduled for release early next year and originally due out at the end of 2004--won't come out at all now. Instead, Intel will boost performance on its chips by increasing the size of the cache, a pool of memory located on the processor for rapid data access. Current mainstream Pentium 4s now have 1MB cache. In the future, these chips will have 2MB of cache, like Intel's Xeon server chips and the "Extreme Edition" Pentium 4s designed for gaming PCs. Intel will continue to come out with Extreme Edition chips by boosting the bus speed and cache size, said Bill Kirby, director of platform marketing at Intel.
Is this Intel's way of saying they can't produce a stable chip at those speeds? -KF

Source: ZDNet (http://news.zdnet.com/2100-9584_22-5409816.html)

[primesuspect]

Or perhaps it's their way of saying "Gee, mebbe meggahurtzs don't matter that much AFTER all :eek2:"

[Thrax]

I'm pretty sure the Centrino, EMT-64 and the Dothan-based Xeons that are in development are evidence to that.

[KingFish]

What adds insult to injury is that the P4 needs high MHz speeds to make their long pipeline design work less inefficiently. Intel sure has made some strange moves lately. If I owned Intel stock I'd be quite nervous at the moment.

KF

[muddocktor]

Yeah, it seems that lately Intel has been shooting themselves in the foot even more often than AMD has been with 90nm A64.

[Leonardo]

The corporate world will catch on: "Hmm, maybe AMD was on to something when they changed their rating system a couple years ago". What waits to be seen though, is how Ma and Pa Kettle and John Doe react in the aisles of Best Buy and CompUSA. Will the new Intel model numbers give them pause to consider AMD? Will the average PC user finally realize that Intel's MHz trumpeting has really been a hollow tune for over two years?

[madmat]

It's funny, I remember reading that they'd already decided to design another way when they started having the heat issues with prescott and now they're just confirming that.
That's ok, if they don't want to make a 4Ghz part I'll make one reach it just to play.

[Jimborae]

Yeah, it seems that lately Intel has been shooting themselves in the foot even more often than AMD has been with 90nm A64.

Mud, what's the problem with the new A64's? I've not read anything & I'm getting a s939 3200 soon, hope I've not made a mistake. :confused:

[Thrax]

They run slightly hotter than their .13u counterparts. The upshot being that it makes room for more cache, SSE3 implementation, a larger DDR2 controller, and dual core, whenever those happen to hit the market.

[madmat]

From everything I've read when AMD goes DDR2 and/or dual core you'll be seeing a new socket introduction anyways...

[Thrax]

DDR2 doesn't require a new socket, literally only a few hundred new transistors, and according to AMD, dual core will be 939-compatible.

[madmat]

It will require a new socket because the CPU has a different memory controller...how else will they make it work, sell 939 CPU's that have big letters stating "FOR DDR2 MOBO'S ONLY!!" or complicate the chip by adding a controller that swings both ways to enable the chips with DDR2 support to be backwards compatible with older chipsets?
If it could even be done it would in all likelyhood reduce the advantage the 64's have over Intel's offerings in terms of sheer memory bandwidth to a lower percentage which would be bad juju for AMD seeings how they have the performance king ATM.
I'll eat my words when a DDR2 939 mobo comes out but I really don't think I'll need to break out the salt and pepper anytime soon.

[mmonnin]

Why would it require a new socket? Did the athlon have to change sockets with the switch from SDRAM to DDR?

[Thrax]

It will require a new socket because the CPU has a different memory controller

That's the problem, Madmat. It won't. DDR2 controllers ALL support DDR1, because the standard was designed that way. The only difference is in the slot architecture between DDR1 and DDR2. Expanding the memory controller on the 939 A64 is not that big a change.

Get the salt and pepper. They won't be making a socket change for DDR2.

[madmat]

Like I said, when I see a 939 DDR2 board I will happily eat my words, only time will tell.

[mmonnin]

You think AMD would introduce 3 sockets with 64-bit goodness and then KNOW that they would have to turn around and make another socket for DDR2? They knew DDR2 was coming a long time ago and I am sure they built some sort of support for the memory change into the socket. They are too smart to do something SOOO stupid.