kirangp
The joy Of Wow
At a press meeting today, Intel's Pat Gelsinger also made a number of high-level disclosures about the successor to Penryn, the 45nm Nehalem core.
Unlike Penryn, which is a shrink/derivative of Core 2 Duo (Merom), Nehalem is architected from the ground up for 45nm. This is a major new design, and Gelsinger revealed some truly tantalizing details about it.
Nehalem has its roots in the four-issue Core 2 Duo architecture, but the direction that it will take Intel is apparent in Gelsinger's insistence that, "we view Nehalem as the first true dynamically scalable microarchitecture." What Gelsinger means by this is that Nehalem is not only designed to take Intel up to eight cores on a single die, but those cores are meant to be mixed and matched with varied amounts of cache and different features in order to produce processors that are tailored to specific market segments.
The blockbuster revelation is that some Nehalem designs will sport an on-die memory controller and integrated graphics processor. Let me talk about the latter before I touch on the former.
A few questioners tried to get clarification from Gelsinger as to whether he meant that there would be a GPU integrated onto the actual die along with the general-purpose CPU cores. (Recall that AMD claims this CPU/GPU die-level integration for their Fusion project.) Gelsinger clarified that the GPU would be "in the socket" with the CPU, but wouldn't say more.
Reading between the lines on this comment and others, I can say with a pretty high degree of certainty Intel will almost certainly be using its packaging skills to put a GPU in the same package as a Nehalem CPU. Furthermore, this is going to help out with mobile products, small-form-factor devices (*cough* Apple), and anywhere else that power and cooling are more critical than raw performance. I'd expect that such CPU/GPU devices will cut down on the number of on-die cores that you can put on the CPU die (for power dissipation reasons). So I expect a Nehalem CPU/GPU combination to consist of a dual-core Nehalem CPU that sports an on-die memory controller, with a GPU sandwiched in the same package with it. This will be a killer mobile part in terms of performance per watt.
As far as the on-die memory controller, this has been a long time coming for Intel. Intel is saying that not every Nehalem part will necessarily have an integrated memory controller—they'll tweak the number of cores and the on-die components to fit certain performance-per-watt targets for certain segments. I expect all of the parts that sport an in-package GPU to also have an on-die memory controller, for performance and power reasons, but I may be wrong.
The other big revelation is that Nehalem will bring back simultaneous multithreading to Intel's processor line. Gelsinger says that Nehalem cores will support up to two threads, and that the implementation will be "like Hyperthreading." There aren't a lot of details yet, but Gelsinger did say that Nehalem parts would max out at eight cores and 16 threads.
Nehalem will go into production in 2008 on Intel's 45nm process.
SOURCE
Unlike Penryn, which is a shrink/derivative of Core 2 Duo (Merom), Nehalem is architected from the ground up for 45nm. This is a major new design, and Gelsinger revealed some truly tantalizing details about it.
Nehalem has its roots in the four-issue Core 2 Duo architecture, but the direction that it will take Intel is apparent in Gelsinger's insistence that, "we view Nehalem as the first true dynamically scalable microarchitecture." What Gelsinger means by this is that Nehalem is not only designed to take Intel up to eight cores on a single die, but those cores are meant to be mixed and matched with varied amounts of cache and different features in order to produce processors that are tailored to specific market segments.
The blockbuster revelation is that some Nehalem designs will sport an on-die memory controller and integrated graphics processor. Let me talk about the latter before I touch on the former.
A few questioners tried to get clarification from Gelsinger as to whether he meant that there would be a GPU integrated onto the actual die along with the general-purpose CPU cores. (Recall that AMD claims this CPU/GPU die-level integration for their Fusion project.) Gelsinger clarified that the GPU would be "in the socket" with the CPU, but wouldn't say more.
Reading between the lines on this comment and others, I can say with a pretty high degree of certainty Intel will almost certainly be using its packaging skills to put a GPU in the same package as a Nehalem CPU. Furthermore, this is going to help out with mobile products, small-form-factor devices (*cough* Apple), and anywhere else that power and cooling are more critical than raw performance. I'd expect that such CPU/GPU devices will cut down on the number of on-die cores that you can put on the CPU die (for power dissipation reasons). So I expect a Nehalem CPU/GPU combination to consist of a dual-core Nehalem CPU that sports an on-die memory controller, with a GPU sandwiched in the same package with it. This will be a killer mobile part in terms of performance per watt.
As far as the on-die memory controller, this has been a long time coming for Intel. Intel is saying that not every Nehalem part will necessarily have an integrated memory controller—they'll tweak the number of cores and the on-die components to fit certain performance-per-watt targets for certain segments. I expect all of the parts that sport an in-package GPU to also have an on-die memory controller, for performance and power reasons, but I may be wrong.
The other big revelation is that Nehalem will bring back simultaneous multithreading to Intel's processor line. Gelsinger says that Nehalem cores will support up to two threads, and that the implementation will be "like Hyperthreading." There aren't a lot of details yet, but Gelsinger did say that Nehalem parts would max out at eight cores and 16 threads.
Nehalem will go into production in 2008 on Intel's 45nm process.
SOURCE