Intel Corp., the world’s largest x86 chipmaker, may have canned its next speed bin of Core 2 Extreme dual-core processor in favour of lower clock-speed quad-core central processing unit.
Intel’s first quad-core microprocessor for desktops is projected to emerge in Q4 2006, a quarter earlier than expected originally, operate at 2.66GHz, use 1066MHz processor system bus and be positioned as Intel’s top-of-the-range offering for gamers and enthusiasts, reports HKEPC web-site, which claims that it had seen the latest roadmaps from Intel Corp.
The new processor will cost $999 in 1000-unit quantities and is likely to substitute the already announced officially 3.20GHz Core 2 Extreme chip with two processing engines. Even though the reasons behind such move are unclear, the transition of “extreme” processor to a multi-core design should emphasize the company’s plan to shift the attention of end-users to the number of cores, not clock-speed.
Given that not all applications can benefit from more than two threads processed at once, the 2.66GHz quad-core microprocessor may offer lower performance in certain applications compared to the 2.93GHz dual-core Intel Core 2 Extreme product.
Earlier it was reported that Kentsfield is a code-name for next-generation “extreme” desktop processor from Intel, will have four processing engines, in contrast to two cores sported by the Core 2 Extreme chips. It is highly likely that Kentsfield features two Conroe chips on the same piece of substrate. Provided that every Conroe chip features two processing engines, the Kentsfield will have four cores, 4MB or 8MB level-two (L2) cache, 1066MHz processor system bus and other characteristics similar to Conroe.
Intel Kentsfield is expected to be drop-in compatible with some of the Intel 975X-based infrastructure that supports Intel Core 2 Duo and Core 2 Extreme processors.
Intel did not comment on the new-story.
Comments currently:
3
Discussion started: 08/19/06 07:47:57 AM
Latest comment: 09/03/06 01:37:06 PM
[1-3]
1.
Fast PRocessor's may no scale. I tis one thing to test hand pick cores and overclock them but it is another to overclock a cpu purchase from a retailer. Conroe is a nice processor but 20% is not awe inspiring. AMD is competing with 90nm against Intel's 65nm and BIG caches an dclocks just as high it will be interesting to see how AMDs clock when they change over to 65nm. It seems funny that when Intel did not perform as well at the same clock speed as a AMD it did not matter but now that things are reversed it seems to matter more. Simply put all AMD has to do is sell a processor at the Same price (or hopefully a bit lower) that is clocked 20-30% quicker than an the intel at the same price. If AMD does this then they are and they are fully competative. I have tried to buy an Conroe E6X00 and other than the low speed 1mb cache version the high locked version are unavailable or VERY expensive. Who cares how fast they are if you cant buy them. If Intel does not get their clock speed up and I mean for real not hand pick review samples, then AMD may just beat by clock their parts higher.
[
Posted by: rjh

|
Date: 08/19/06 07:47:57 AM]
2.
Their processors scale reasonably well. Most 2.4-2.67 users are getting between 3.2 and 3.5ghz out of those chips and exteme chips 2.93's are doing in the 3.5-3.8 range. That's not entirely a slow chip. So yes let's take the A64 and say clock it to what 4-4.5ghz to compete?
That's gonna take a while...
AMD's answers more likely lie in the K8L, not in a shrink of the current design alone.
[
Posted by: FXi

|
Date: 08/20/06 11:12:23 PM]
3.
mother fuckers
[
Posted by: matiasg

|
Date: 09/03/06 01:37:06 PM]
[1-3]