"Anyone got much more of a clue about when we'll see the 90nm Athlon 64?"
Roadmap clearly states second half 2005, which is *at least* 12-18 months from now. With the usualy roadmap optimism deducted, don't expect them until 2006. So you go buy some other CPU meanwhile. Always look at what you can buy now and what platform is more future proof, never sit and wait.
Dual cores will have local access to RAM, which is good. On a dual Opteron today, on a non-NUMA aware OS, half the memory accesses for one CPU goes across the second CPU. Dual Opteron cores will therefore perform better than dual CPUs with no NUMA, but with NUMA, data is placed locally on each CPU and performance gets more even. Dunno which one wins overall, though.
On a dual Xeon, it doesn't matter as much, both single CPUs and dual cores in a dual core CPU will go past the NB anyway. Dual cores means sharing the bandwidth to the NB, just like in a dual Xeon today. Intels buses all share the FSB to the NB, always has. (Even AXP supported separate FSBs for each CPU to the NB, pity they never made nForce2 dual DDR for dual AXPs)
Fred Weber said in an interview that dual core Opterons support 4 way. This is logical as the core will still need to talk L2 cache coherency via the HT protocols and there are only 3 bits for core IDs.
An I missing something here, but, as the rest of you already figured out, the article gives no reference to "Toledo" on the roadmap, which is dual core for desktop.
To me this, this is the biggest news. We already knew servers would be going dual core; now we know it will quickly find its way to desktop.
Ok well 20 emails later, ok they are promising it. That's a good thing!! In spite of me eating crow :(
Concentrate on that being a good thing because I think they needed to do it very much, give dual core on at least the FX line. I kinda doubt it will happen in 2005, but I won't let that detract from it being a good move.
It's worth reiterating that this is OPTERON chips. AMD has stated several times that it has NO plans at the moment to bring dual cores to the desktop. Since you can well imagine what a double FX chip would cost I think you can easily see why...
So yep exciting news, but none of this is for a 939 socket, even with a single memory controller.
"AMD's dual-core server processors will share a single memory controller, Weber said. This won't create a bottleneck because a server with two Opteron chips, and therefore two memory controllers, already has more than enough memory bandwidth required to run that system, he said."
I'm still wondering about that P4 720 3.73Gig and 2Mb of cache, lumped in with the banias and dothans in the roadmap, that's supposed to be coming out before year's end. I get the feeling that both companies are pushing their dual core R&D at highest priority.
If I remember right, I saw many stories around the web saying every opteron's memory controller had two imput ports to the memory controller from the CPU. Thus, you can go dual core and NOT need to redesign the memory controller at all, and you wont need to change the pinout on the socket. Thus you could take out a single core S940 opteron and put in a dual core opteron. Power issues are the only thing complicating it I believe.
I suspect the Opteron 8xx's can do a lot more than 16-way processing. Opteron 8xx has 3 HyperTransport ports that can connect to 3 other CPUs. If the HT ports are connected differently from what was suggested by AMD, it can theoretically run an infinitely long chain of processors.
As interesting as 2x128-bit memory controllers are, having to install 4 memory modules is a little excessive to most, and most likely requires registered DIMM or Intel's FB-DIMM to work. Hopefully there will be 256/288-bit memory modules when dual core Opteron arrives.
dang, 16 way proccessing on one board? I assume i'm reading the roadmap right. Saying the 1xx series, dual core, is actually 2 proccessors. That would mean the 8xx series supports 16. Just think what a Rackable stand with 80 servers, each with 16x854's Optrons can do. 1280 Optrons working an parallel on one rack. Holy Cow.
It would be, but I imagine AMD might save that as their trump card if they really do need it, especially since the whole S940/939 fiasco already managed to annoy people, and another socket change would go even furthur in that.
Are those all implications? If AMD does decide to activate both memory controllers, it would have big advantages in UMA-enabled operating systems like Longhorn will be. You'd have twice dual-channel memory bandwidth in that situation, which is potentially a big advantage compared to Intel's dual-core offerings.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
20 Comments
Back to Article
Degrador - Tuesday, June 15, 2004 - link
Yeah, I do - socket 939 is coming and I figure I may as well start that with a 90nm processor.DerekBaker - Tuesday, June 15, 2004 - link
Pjotr, I think he meant the single-cored models. Which means the second half of this year.Derek
Pjotr - Tuesday, June 15, 2004 - link
"Anyone got much more of a clue about when we'll see the 90nm Athlon 64?"Roadmap clearly states second half 2005, which is *at least* 12-18 months from now. With the usualy roadmap optimism deducted, don't expect them until 2006. So you go buy some other CPU meanwhile. Always look at what you can buy now and what platform is more future proof, never sit and wait.
Pjotr - Tuesday, June 15, 2004 - link
dual core vs dual CPUs?Dual cores will have local access to RAM, which is good. On a dual Opteron today, on a non-NUMA aware OS, half the memory accesses for one CPU goes across the second CPU. Dual Opteron cores will therefore perform better than dual CPUs with no NUMA, but with NUMA, data is placed locally on each CPU and performance gets more even. Dunno which one wins overall, though.
On a dual Xeon, it doesn't matter as much, both single CPUs and dual cores in a dual core CPU will go past the NB anyway. Dual cores means sharing the bandwidth to the NB, just like in a dual Xeon today. Intels buses all share the FSB to the NB, always has. (Even AXP supported separate FSBs for each CPU to the NB, pity they never made nForce2 dual DDR for dual AXPs)
Pjotr - Tuesday, June 15, 2004 - link
Fred Weber said in an interview that dual core Opterons support 4 way. This is logical as the core will still need to talk L2 cache coherency via the HT protocols and there are only 3 bits for core IDs.Degrador - Tuesday, June 15, 2004 - link
Anyone got much more of a clue about when we'll see the 90nm Athlon 64? I'm looking to upgrade soon and am thinking of the overclocking potential :)mjrpes2 - Monday, June 14, 2004 - link
Errr, nevermind. I'm too sleep deprived.mjrpes2 - Monday, June 14, 2004 - link
An I missing something here, but, as the rest of you already figured out, the article gives no reference to "Toledo" on the roadmap, which is dual core for desktop.To me this, this is the biggest news. We already knew servers would be going dual core; now we know it will quickly find its way to desktop.
Yupee! :)
skunkbuster - Monday, June 14, 2004 - link
whats the difference between dual core and having dual processors? is there any performance advantage in having dual core?Anemone - Monday, June 14, 2004 - link
Ok well 20 emails later, ok they are promising it. That's a good thing!! In spite of me eating crow :(Concentrate on that being a good thing because I think they needed to do it very much, give dual core on at least the FX line. I kinda doubt it will happen in 2005, but I won't let that detract from it being a good move.
Anemone - Monday, June 14, 2004 - link
It's worth reiterating that this is OPTERON chips. AMD has stated several times that it has NO plans at the moment to bring dual cores to the desktop. Since you can well imagine what a double FX chip would cost I think you can easily see why...So yep exciting news, but none of this is for a 939 socket, even with a single memory controller.
eRacer - Monday, June 14, 2004 - link
AMD confirmed in the Infoworld article below that the dual core CPUs will only have one memory controller.http://www.infoworld.com/article/04/06/14/HNamddua...
"AMD's dual-core server processors will share a single memory controller, Weber said. This won't create a bottleneck because a server with two Opteron chips, and therefore two memory controllers, already has more than enough memory bandwidth required to run that system, he said."
Doormat - Monday, June 14, 2004 - link
The article is here...http://www.theinquirer.net/?article=13344
Pumpkinierre - Monday, June 14, 2004 - link
I'm still wondering about that P4 720 3.73Gig and 2Mb of cache, lumped in with the banias and dothans in the roadmap, that's supposed to be coming out before year's end. I get the feeling that both companies are pushing their dual core R&D at highest priority.Doormat - Monday, June 14, 2004 - link
If I remember right, I saw many stories around the web saying every opteron's memory controller had two imput ports to the memory controller from the CPU. Thus, you can go dual core and NOT need to redesign the memory controller at all, and you wont need to change the pinout on the socket. Thus you could take out a single core S940 opteron and put in a dual core opteron. Power issues are the only thing complicating it I believe.quanta - Monday, June 14, 2004 - link
I suspect the Opteron 8xx's can do a lot more than 16-way processing. Opteron 8xx has 3 HyperTransport ports that can connect to 3 other CPUs. If the HT ports are connected differently from what was suggested by AMD, it can theoretically run an infinitely long chain of processors.As interesting as 2x128-bit memory controllers are, having to install 4 memory modules is a little excessive to most, and most likely requires registered DIMM or Intel's FB-DIMM to work. Hopefully there will be 256/288-bit memory modules when dual core Opteron arrives.
AtaStrumf - Monday, June 14, 2004 - link
It could get the No.1 spot in folding in about 2 days :)Runamile - Monday, June 14, 2004 - link
dang, 16 way proccessing on one board? I assume i'm reading the roadmap right. Saying the 1xx series, dual core, is actually 2 proccessors. That would mean the 8xx series supports 16. Just think what a Rackable stand with 80 servers, each with 16x854's Optrons can do. 1280 Optrons working an parallel on one rack. Holy Cow.GokieKS - Monday, June 14, 2004 - link
It would be, but I imagine AMD might save that as their trump card if they really do need it, especially since the whole S940/939 fiasco already managed to annoy people, and another socket change would go even furthur in that.~KS
Nighteye2 - Monday, June 14, 2004 - link
Are those all implications? If AMD does decide to activate both memory controllers, it would have big advantages in UMA-enabled operating systems like Longhorn will be. You'd have twice dual-channel memory bandwidth in that situation, which is potentially a big advantage compared to Intel's dual-core offerings.