Intel's Dual Core Strategy Investigated
by Anand Lal Shimpi on October 22, 2004 3:09 PM EST- Posted in
- CPUs
The Problem with Intel's Approach
The major issue with Intel's approach to dual core designs is that the dual cores must contest with one another for bandwidth across Intel's 64-bit NetBurst FSB. To make matters worse, the x-series line of dual core CPUs are currently only slated for use with an 800MHz FSB, instead of Intel's soon to be announced 1066MHz FSB. The reduction in bandwidth will hurt performance scalability and we continue to wonder why Intel is reluctant to transition more of their CPUs to the 1066MHz FSB, especially the dual core chips that definitely need it.
With only a 64-bit FSB running at 800MHz, a single x40 processor will only have 6.4GB/s of bandwidth to the rest of the system. Now that 6.4GB/s is fine for a single CPU, but an x40 with two cores the bandwidth requirements go up significantly.
AMD's Strategy
While Intel's current roadmap appears to place dual core on the desktop before it makes its way to the enterprise (other than with Itanium), AMD's strategy is reversed - with dual core appearing in workstations and on servers before making a splash on the desktop.
Overall, AMD's approach simply makes more sense, since the overall performance benefit to dual core on the desktop will be minimal at best but strong in very specific applications and usage patterns. With most desktop applications continuing to be single threaded, dual core will still have to wait until there is more application support before truly being useful on the desktop. Heavy multitaskers and those running workstation applications will appreciate the benefits of dual core, but gamers and most other users will find higher clocked single core chips to be better suited for their needs.
The scenario is exactly the opposite in the workstation and server space, with the applications already seeing huge benefits from going to multiple processors thanks to their multithreaded nature.
When AMD mentions that their K8 architecture was designed for multicore operation from the start, they weren't lying. Each Socket-939 or Socket-940 K8 chip, whether it's an Athlon 64, Athlon 64 FX or Opteron, features three Hyper Transport links (whether they are all operational is another question). In order to create a dual core version of a K8 based chip, you simply remove a single pair of Hyper Transport PHYs, one from each chip, and fuse the two Hyper Transport links together - thus creating a direct path of communication between the two cores, capable of transmitting data at up to 8GB/s (at 1GHz) between the two chips. Update: There is some debate as to how AMD implements dual core in their K8 architecture. The above description was provided by AMD from an earlier discussion but many readers have emailed to point out that the two cores are connected at the SRQ level. We are awaiting official confirmation from AMD as to exactly how their dual core technology is implemented. Update 2:While AMD never got back to us with an official response, unofficially they did confirm that the two cores on a single dual core Opteron die do communicate at full speed and are not connected at the HT level. We apologize for the error.
AMD's performance limitation here will be memory bandwidth, with the two K8 cores sharing the 128-bit DDR memory bus. While we currently don't see a huge performance increase from going to a 128-bit memory bus from a single channel 64-bit interface, the move to dual core will definitely make greater use of memory bandwidth.
AMD continues to list the second half of 2005 as the introduction timeframe for their dual core CPUs, with Opteron coming first and then Athlon 64 FX. Once again, as with all release dates, nothing is set in stone, but right now it looks like that both AMD and Intel are planning on having dual core on the desktop in the same general timeframe.
AMD has yet to reveal what the official specifications of their upcoming dual core desktop products are, but based on roadmaps and what we've seen, it would seem that the first dual core desktop parts will be based on two 90nm Athlon 64 FX cores with a shared memory controller. Interally AMD is referring to this CPU as "Toledo" as we've already published.
Facebook
Twitter
RSS
59 Comments
View All Comments
ViRGE - Saturday, October 23, 2004 - link
#35, all the compilers in the world won't help Intel here with multicore. Threading isn't just an optimization, it's something that has to be explicitly declared and controlled in code in order to make sure that the threads don't work against each other, or assume that the other is/has done something it hasn't. Intel may be able to work in some thread-swapping tricks to speed things up, but for multicore designs to shine, it's up to the developers to do it; Intel can't control that.Anemone - Saturday, October 23, 2004 - link
Actually Intel has confirmed again and again that dual core is coming in Q3 05. Those are netburst cpu's so that's some approximation of dual P4. Those will plug into current LGA775, and run at 2.8, 3.0 and 3.2, with 64 bit ability.Chuckles - Saturday, October 23, 2004 - link
I'm surprised that games haven't been multi-threaded on the PC side. The games for Mac are...michaelpatrick33 - Saturday, October 23, 2004 - link
#35 Intel itself has that the dual core P4's won't be out until Q1 2006. That is 2006. They are delaying the Celeron D 350 and the Pentium 4 670 until Q2 05 and they will be on a 800 FSB. The only dual core Intel is releasing in 2005 is the Itanium. They will release a dualcore P4 chipset in 2005 but have no dualcore P4's to go into it until 2006. These aren't rumors, these are the facts. Intel is only going to take a small hit financially but they are definitely behind AMD on some fronts at this time. This doesn't mean AMD won't shoot itself in the foot but Intel is facing some problems. There have been a number of 64bit beta programs that show tremendous gains in 64bit long mode due more to the 16gpr's than 64bit. The 64bit part of AMD x86-64 is more than just 64bit remember. So is Intel's. Windows XP64 will be out 1st half 05 (supposedly) and we will see 64bit programs along with it. Should be interestingAnemone - Friday, October 22, 2004 - link
No no no and no.I'll forgive the bias, but if this is your honest summary opinion of whare Intel is headed, or what problems will be its Achilles heal (memory bandwidth) I can only say that I think the coming year is going to be quite an education for you.
Don't get that entirely wrong. AMD is most definitely doing things very well, but their plan to introduce dual core to servers first is based on technological limitations, the markets they feel they still need to make inroads on (strategic order), and where they will achieve the best margins. They will still have to share a bus, and still have to pass cache duplication over that same bus, so I'd be highly careful on just what you think their chips will do in dual core format. Good design is probable, but AMD can and has let us down more than once.
Heck, so has Intel. Prescott? God, what WERE they thinking? It has improved, but 40 I mean 64 bit is still like mystery soup. I'm positive they just want to sell everyone new "64 bit" Prescotts in 2005, so they're robbing every customer now, imo. But don't think for one minute that they've taken this year as it stands without working behind the scenes. As for forcing dual core and then letting the market make software for it? Well let's see, MMX, SSE, SSE2, doesn't that seem like a trend? And would you say that SSE2 is useless, and unsupported? Intel builds the compilers that make their hardware run. They even make the compilers that make AMD hardware run. That's pretty key, and rather silly to ignore.
I guess when I read this I hadn't expected you to look at every possible Intel rumour and take the "worst possible scenario" route to find your best guess answer. I was wrong.
cosmotic - Friday, October 22, 2004 - link
Why no word about Moterola or IBM? Granted they aren't PC chips, it would be nice to see a comparison.AlphaFox - Friday, October 22, 2004 - link
#19 - win 2000 doesnt know the difference between a dual processor system and a hyper threaded cpu. win XP and server 03 know that it is just a single CPU with diper threading enabled, thus you can run a dual processor hyperthreaded system on XP pro.ViRGE - Friday, October 22, 2004 - link
#31, it's similar to, but not the same as the 64bit transition. At the very least, if you want to get some use out of the 64bit extras, you simply need to recompile your program, and you can work out further optimizations without too much effort. Converting programs to usefully use multicore tech though is going to be far more difficult, since it requires a concept(multithreading) that isn't known very well by most programmers, implemented even less, and no one is really sure what a good way is to multithread some types of applications(i.e. games).xsilver - Friday, October 22, 2004 - link
Hang on -- to all those that say dual threads are crap -- what exactly are you running -- AMD 64 maybe? they'res no software that can take advantage of the 64 bit, so its virtually the same thing no?and to those that say the heat requirements will stay the same because of increased die size, in total isn't the while cpu still socket 775..... that's still pretty small and probably too much for air cooling to handle. Some sites report a 30 degree increase at load for 100W of power (on a thermalright heatpipe i think) 200w will be way too high in temp increase?
michaelpatrick33 - Friday, October 22, 2004 - link
Intel has said their desktop dual coares won't come out until Q1 2006 and that the only dual cores coming out in 2005 was their Itanium.http://www.xbitlabs.com/news/cpu/display/200410200...