Nehalem due in Q4 2008
http://www.dailytech.com/article.aspx?newsid=9417
Now doesn't THIS sound tasty? An integrated memory controller, SMT and L3 cache designs. I've heard that there will be versions with GPU integrated into the die as well. I'm betting that the OS X 10.6 design is going to be heavily influenced by Nehalem to good effect.
Quote:
New sockets, chipsets and architecture en route from Intel before 2009
Nehalem will likely be the most aggressive processor architecture in Intel's portfolio since the original Pentium. With the launch of the Core architecture, the company announced its tick-tock strategy: design new architecture, then shrink the process node. Rinse and repeat.
Tick-tock is alive and well as Intel's corporate roadmap reveals additional details about its desktop iteration of 45nm quad-core Nehalem, dubbed Bloomfield.
Nehalem will be fundamentally different from the Core architecture for no less than two reasons. The company will move the memory controller from the core logic on the motherboard to the processor die. This tactic has been a cornerstone for the AMD K8 architecture since 2003.
In addition, Nehalem will also feature a new bus interconnect, currently dubbed Quick Path Interconnect. This new interconnect behaves very similar to HyperTransport, currently used on all AMD platforms since K8.
A new bus and memory controller means a new socket design. Existing motherboards are not compatible with Nehalem-based processors. The new desktop socket, labeled LGA1366, will completely replace the existing LGA775 interconnect.
The company will replace the X38 and yet to be announced X48 desktop chipsets with the Tylersburg chipset family and ICH10 southbridge for these first LGA1366 motherboards.
Corporate guidance also suggests the company will likely ditch all DDR2 support in favor of DDR3, at least on the high end platforms. All Bloomfield processors will feature support three DDR3 channels.
However, not everything is known about Nehalem just yet. Corporate guidance suggests Bloomfield will feature a new revision of Hyper-Threading. Although each Bloomfield features four physical cores, the processor will dynamically allocate additional threads -- Bloomfield computers will detect eight logical cores.
Bloomfield will feature less cache than Intel's high-end 45nm Penryn offerings slated for release between now and Q4 2008. However, unlike the 12MB L2 cache featured on Penryn, the 8MB L3 cache on all Nehalem offerings can be shared between all four on-die cores.
Intel's highest-end Bloomfield processors will feature a 130W thermal envelope. Extreme Edition Penryn processors, the first on the 45nm node, have a thermal envelope that tops out around 136W. Intel's Q9550 processor (2.8 GHz, 45nm quad-core) sports a 95W TDP.
New sockets, chipsets and architecture en route from Intel before 2009
Nehalem will likely be the most aggressive processor architecture in Intel's portfolio since the original Pentium. With the launch of the Core architecture, the company announced its tick-tock strategy: design new architecture, then shrink the process node. Rinse and repeat.
Tick-tock is alive and well as Intel's corporate roadmap reveals additional details about its desktop iteration of 45nm quad-core Nehalem, dubbed Bloomfield.
Nehalem will be fundamentally different from the Core architecture for no less than two reasons. The company will move the memory controller from the core logic on the motherboard to the processor die. This tactic has been a cornerstone for the AMD K8 architecture since 2003.
In addition, Nehalem will also feature a new bus interconnect, currently dubbed Quick Path Interconnect. This new interconnect behaves very similar to HyperTransport, currently used on all AMD platforms since K8.
A new bus and memory controller means a new socket design. Existing motherboards are not compatible with Nehalem-based processors. The new desktop socket, labeled LGA1366, will completely replace the existing LGA775 interconnect.
The company will replace the X38 and yet to be announced X48 desktop chipsets with the Tylersburg chipset family and ICH10 southbridge for these first LGA1366 motherboards.
Corporate guidance also suggests the company will likely ditch all DDR2 support in favor of DDR3, at least on the high end platforms. All Bloomfield processors will feature support three DDR3 channels.
However, not everything is known about Nehalem just yet. Corporate guidance suggests Bloomfield will feature a new revision of Hyper-Threading. Although each Bloomfield features four physical cores, the processor will dynamically allocate additional threads -- Bloomfield computers will detect eight logical cores.
Bloomfield will feature less cache than Intel's high-end 45nm Penryn offerings slated for release between now and Q4 2008. However, unlike the 12MB L2 cache featured on Penryn, the 8MB L3 cache on all Nehalem offerings can be shared between all four on-die cores.
Intel's highest-end Bloomfield processors will feature a 130W thermal envelope. Extreme Edition Penryn processors, the first on the 45nm node, have a thermal envelope that tops out around 136W. Intel's Q9550 processor (2.8 GHz, 45nm quad-core) sports a 95W TDP.
Now doesn't THIS sound tasty? An integrated memory controller, SMT and L3 cache designs. I've heard that there will be versions with GPU integrated into the die as well. I'm betting that the OS X 10.6 design is going to be heavily influenced by Nehalem to good effect.
Comments
If plans from Advanced Micro Devices and Intel bear fruit, the next big bump in computer performance may come from the descendants of today's video cards.
Many companies have begun viewing the graphics processing unit, or GPU -- the chip found on video cards -- as a significant source of computing power in its own right. For instance, GPUs are now being used to tackle computing tasks like bioinformatics, cryptography and audio-signal processing, not just to render visual images onscreen.
AMD and Intel want to take this trend a step further and incorporate a GPU on the same chip as the central processing unit (CPU), possibly by late 2008 or early 2009.
I'm just interested in the price savings vs performance/watt. Will ondie GPU outperform Integrated Graphics in 2009? We'll see.
I'm just interested in the price savings vs performance/watt. Will ondie GPU outperform Integrated Graphics in 2009? We'll see.
There will also be some very interesting form-factor changes. If the GPU effectively moves on die with the CPU, that frees up some space and reduces the overall heat signature allowing for smaller form-factors.
And when you couple that with the likelihood of flash drives and other shrinkages: How about a Mini-sized Mac Pro?
There will also be some very interesting form-factor changes. If the GPU effectively moves on die with the CPU, that frees up some space and reduces the overall heat signature allowing for smaller form-factors.
And when you couple that with the likelihood of flash drives and other shrinkages: How about a Mini-sized Mac Pro?
Apple could surely move in this direction with a Mac Pro the size of say a Shuttle PC.
Imagine this
Cube like form factor
Nehalem 4-Core with shared L3 and Integrated GPU
2GB of DDR3
DVD/HD DVD/Blu-ray Universal drive.
Three 2.5" Drive Bays
10G Ethernet, Two FW800 busses and 1 FW400
PCI Express 2.0 16x and 4x PCIe slots
HDMI output
Wireless A/B/G/N
OS X 10.5.3 (Resolution Independence, AACS/BD+ enabled)
$1499
Now imagine how the 10G Ethernet could be used to tie these computers together in a XGrid array? HPC out of nowhere.
I'm not downplaying the forthcoming Penryn based Macs but a year from now we're looking at a new Intel architecture which is more saliva inducing IMO.
Apple could surely move in this direction with a Mac Pro the size of say a Shuttle PC.
Imagine this
Cube like form factor
Nehalem 4-Core with shared L3 and Integrated GPU
2GB of DDR3
DVD/HD DVD/Blu-ray Universal drive.
Three 2.5" Drive Bays
10G Ethernet, Two FW800 busses and 1 FW400
PCI Express 2.0 16x and 4x PCIe slots
HDMI output
Wireless A/B/G/N
OS X 10.5.3 (Resolution Independence, AACS/BD+ enabled)
$1499
Now imagine how the 10G Ethernet could be used to tie these computers together in a XGrid array? HPC out of nowhere.
I'm not downplaying the forthcoming Penryn based Macs but a year from now we're looking at a new Intel architecture which is more saliva inducing IMO.
I agree that Nehalem looks more groundbreaking than Penryn does, but Penryn will get my money anyway. I need a new box in the next couple of months and can't afford to wait for Nehalem.
I have a feeling that we will see the day when the size of the peripheral connectors are a factor in determining the size of the computer.
I'm not downplaying the forthcoming Penryn based Macs but a year from now we're looking at a new Intel architecture which is more saliva inducing IMO.
Hell you could start a new thread when debating Intel's tick tock strategy, 'what's the greater impact, die shrink or architecture change'.
Hard to say at this point as Intel seems to have winner with just about everything they release lately.
I agree that Nehalem looks more groundbreaking than Penryn does, but Penryn will get my money anyway. I need a new box in the next couple of months and can't afford to wait for Nehalem.
I have a feeling that we will see the day when the size of the peripheral connectors are a factor in determining the size of the computer.
I wouldn't wait either. If you need a fast computer now Penryn is going to be very nice. We need to see how good Leopard's NSOperation and NSOperationQueue is regaring dealing with multi-core Macs and setting prioritities.
Nahalem is going to requires very smart threading and prioritization. Even a base quad core Nehalem computer is going to appear to the OS and above as a 8 Logical cores. Apple hasn't really maximized performance on more than 4-cores but Leopard may finally change things.
Since this is future hardware I think all discussion regarding Penryn and Nehalem is appropriate as both are shipping in the next 12 months.
Hell you could start a new thread when debating Intel's tick tock strategy, 'what's the greater impact, die shrink or architecture change'.
Hard to say at this point as Intel seems to have winner with just about everything they release lately.
My vote would be architectural change. The move to .90 nano tech was supposed to follow typical trends for die shrinks yet IBM, Intel and AMD all struggled with .90nm parts. That bath of fire probably helped Intel get 65 nano tech working properly and now they seem pretty confident that 45 and then 32 nano tech is going to be relatively straight forward.
Architectural changes seem to deliver more bang for the buck IMO. From ondie memory controllers to more efficient cache technology the architecture lays the groundwork for software to catch up. The reason why Parallels and Fusion work so well isn't merely because the CPU is getting faster but rather the Intel hardware has explicit support for virtualizing software (which is improving in performance with the Penryn supporting chipsets).
My recommendation is always to go with what works for the current but keep an eye out for the future.
My vote would be architectural change. The move to .90 nano tech was supposed to follow typical trends for die shrinks yet IBM, Intel and AMD all struggled with .90nm parts. That bath of fire probably helped Intel get 65 nano tech working properly and now they seem pretty confident that 45 and then 32 nano tech is going to be relatively straight forward.
Architectural changes seem to deliver more bang for the buck IMO. From ondie memory controllers to more efficient cache technology the architecture lays the groundwork for software to catch up. The reason why Parallels and Fusion work so well isn't merely because the CPU is getting faster but rather the Intel hardware has explicit support for virtualizing software (which is improving in performance with the Penryn supporting chipsets).
My recommendation is always to go with what works for the current but keep an eye out for the future.
Yeah Nahalem looks pretty impressive. Could be a real dagger in the heart of AMD as they loose their only advantage over Intel processor design.
I probably will be looking to get a new Mac about then too.
We need to see how good Leopard's NSOperation and NSOperationQueue is regaring dealing with multi-core Macs and setting prioritities.
Well, its the right approach so I think we'll see much success for software that leverages it properly.
Since this is future hardware I think all discussion regarding Penryn and Nehalem is appropriate as both are shipping in the next 12 months.
Agreed. As was the reference to the forth coming 32nm tech. Like you said, this is the future tech forum.
Yeah Nahalem looks pretty impressive. Could be a real dagger in the heart of AMD as they loose their only advantage over Intel processor design.
Let's hope not. Intel needs AMD. Without the competition, Intel was slow and cumbersome and we never would have seen something as aggressive as their "tick-tock" strategy.
I hope AMD comes out with something incredible that keeps Intel on their toes.
P.S. By the same logic I think we will start to see Microsoft be more innovative now that their virtual monopoly is under attack. And that is a good thing for all of us.
I think we will start to see Microsoft be more innovative now that their virtual monopoly is under attack. And that is a good thing for all of us.
The only innovation I can recall coming from Microsoft was taking public domain software and taking it private, which pissed off Richard Stallman with the GPL as a result. Microsoft do, however, make good hardware. They should give up on software and focus on what they're good at -- keyboards and mice.
http://www.dailytech.com/article.aspx?newsid=9417
Now doesn't THIS sound tasty? An integrated memory controller, SMT and L3 cache designs. I've heard that there will be versions with GPU integrated into the die as well. I'm betting that the OS X 10.6 design is going to be heavily influenced by Nehalem to good effect.
Except for the single core CPUs, most new Nehalem CPUs, will get all the new features: single die, QPI Link, Share cache, DRAM controller. It's a new era.
There will also be some very interesting form-factor changes. If the GPU effectively moves on die with the CPU, that frees up some space and reduces the overall heat signature allowing for smaller form-factors.
And when you couple that with the likelihood of flash drives and other shrinkages: How about a Mini-sized Mac Pro?
In an earlier thread, I joked that since Jobs liked the AIO form so much, he might use it for the Mac Pro.
Some things said in jest sometimes come to pass. However, an AIO Mac Pro would HAVE to be easy to upgrade, so it couldn't be in one piece like the iMac, and it would have to be a lot thicker. Apple would have only two "desktop" configurations, and they would look alike.
Remember, you heard it first here.
There will also be some very interesting form-factor changes. If the GPU effectively moves on die with the CPU, that frees up some space and reduces the overall heat signature allowing for smaller form-factors.
Not really. A low-end GPU will move on-die with a low-end CPU, but the high-end GPUs will still be double-wide, a high-end CPU will still have its big heatsink, and a workstation will still have two CPUs. So Macs will stay the same size, but a future Mini will be as powerful as today's Mac Pro, and the future Mac Pro will be much faster than the future Mini.
(And yes, people will need the future Mac Pro to edit their holographic 4K home movies with 11.1 sound.)
The Apple TV might get smaller, although it's limited by the size of the ports. Maybe the inside of the box will be mostly empty, like the Vudu.
Let's hope not. Intel needs AMD. Without the competition, Intel was slow and cumbersome and we never would have seen something as aggressive as their "tick-tock" strategy.
I hope AMD comes out with something incredible that keeps Intel on their toes.
Don't get me wrong, I'm not hoping for AMDs demise, but there just isn't much good news coming out of AMD. I'm just not reading a lot of positive things about Barcelona and Phenom. Whereas good things seem to be coming from Penryn.
Don't get me wrong, I'm not hoping for AMDs demise, but there just isn't much good news coming out of AMD. I'm just not reading a lot of positive things about Barcelona and Phenom. Whereas good things seem to be coming from Penryn.
AMD's Phenom is, I believe, their next big thing. I don't know if it is going to keep them in the ballpark or not. I have to admit that with Apple exclusively using Intel parts, I don't pay much attention to AMD anymore.
AMD's Phenom is, I believe, their next big thing. I don't know if it is going to keep them in the ballpark or not. I have to admit that with Apple exclusively using Intel parts, I don't pay much attention to AMD anymore.
The problem I keep reading about phenom/barcelona is they don't clock high enough. They are going to release a 3 core chip that may be interesting but it seems like they are only going to be competitive at the mid and low end of processor performance. They just don't seem to be able to make a processor to compete with Intel's best at the moment.
That seems to be what's killing AMD because they can't withstand a price war with Intel.