Q3'05 AMD Roadmap; DDR2 and New Sockets for AMD
by Kristopher Kubicki & Jarred Walton on July 21, 2005 1:47 PM EST- Posted in
- CPUs
Server and Workstation Processors
Not much has changed in the server and workstation segment other than the actual launch and availability of the dual core Opterons. Here's what we have looking at the present and near future.
| AMD Server/Workstation Roadmap | ||||
| Processor | Core Name | Clock Speed | Socket | Launch Date |
| Opteron 875 | Egypt | 2.2 GHz 1MB | Socket 940 | Now |
| Opteron 870 | Egypt | 2.0 GHz 1MB | Socket 940 | Now |
| Opteron 865 | Egypt | 1.8 GHz 1MB | Socket 940 | Now |
| Opteron 865 HE | Egypt | 1.8 GHz 1MB | Socket 940 | Now |
| Opteron 860 HE | Egypt | 1.6 GHz 1MB | Socket 940 | Now |
| Opteron 854 | Athens | 2.8 GHz 1MB | Socket 940 | Q3'05 |
| Opteron 275 | Italy | 2.2 GHz 1MB | Socket 940 | Now |
| Opteron 270 | Italy | 2.0 GHz 1MB | Socket 940 | Now |
| Opteron 265 | Italy | 1.8 GHz 1MB | Socket 940 | Now |
| Opteron 265 HE | Italy | 1.8 GHz 1MB | Socket 940 | Now |
| Opteron 260 HE | Italy | 1.6 GHz 1MB | Socket 940 | Now |
| Opteron 254 | Troy | 2.8 GHz 1MB | Socket 940 | Q3'05 |
| Opteron 175 | Denmark | 2.2 GHz 1MB | Socket 939 | Now |
| Opteron 170 | Denmark | 2.0 GHz 1MB | Socket 939 | Now |
| Opteron 165 | Denmark | 1.8 GHz 1MB | Socket 939 | Now |
| Opteron 154 | San Diego | 2.8 GHz 1MB | Socket 939 | Q3'05 |
| Opteron x52 | Venus/Troy/Athens | 2.6 GHz 1MB | Socket 940 | Now |
We've separated the 8xx, 2xx, and 1xx Opteron models into their individual parts for a reason. While older Opterons all targeted socket 940, AMD is apparently moving the 1xx models towards socket 939. The most likely reason is for price/performance advantages of the platform, or perhaps that yields of the newer parts have reached the point where everything works in two socket or higher configurations. 152 is the last single socket Opteron appearing on the roadmap, though determined single socket 940 holdouts can always drop in the more expensive 2xx parts if they want.
Looking at the future, the only new Opterons coming out will be the single core x54 models. As we just mentioned, 154 will actually be a socket 939 part (along with the now available 165, 170, and 175). The switch to socket 939 also means that the 154 will use the San Diego core rather than the Venus core that previous 1xx 90nm parts have been based off. 254 and 854 will also exist, and we must admit we're a little surprised - several of us were surprised to see the x52 models. Almost any application that can take advantage of SMP will perform better on two 265 cores than on two 254 cores, but AMD apparently has received enough requests for another single core Opteron model that the x54 parts are being made. The price of the 254/854 parts is identical to that of the 265/865 parts, however, so the parts are essentially being rated as equivalent. Users that have applications that depend more on pure clock speed rather than multiple threads are the likely target of the x54 parts. The socket 939 154 part is priced the same as the FX-57, and actually costs more than the 254, so single socket 940 users shouldn't fret too much. The 939 152 is also priced the same as the FX-55, while the 940 152 is substantially cheaper.
Besides the various Opteron models most of you have probably seen before, we also included the HE models. HE stands for "High Efficiency" and the efficiency we're talking about is performance per Watt. Where the standard Opterons have a TDP of 95W, the HE models are only 55W. You basically give up two bins of performance for the lower TDP, however, so the 865 HE costs as much as the 875 and the 850 HE costs as much as the 865. They also max out at two levels below the fastest models, which isn't too surprising.
Looking to the future of Opterons, the current roadmap doesn't contain any information about what's planned. DDR2 is in the works for all the other markets, so it would make sense for Opteron to eventually move that way as well. However, there have been difficulties in the past with getting ECC and registered memory to work with DDR2, so perhaps Opterons will move to FB-DIMMs (Fully Buffered DIMMS) instead. We do know that AMD has something in the works called socket F, a 1207 pin socket for future Opterons, but we don't know what type of RAM is used. That transition is likely more than a year off, but we'll let you know as soon as we get any clearer picture of what AMD has planned.
Final Thoughts
Echoing what we said in our last AMD roadmap, there aren't a whole lot of speed increases showing up on the future roadmaps. FX-59 will add another 200 MHz to the fastest AMD processor in terms of clock speed, and we can guess that the X2 5000+ will do the same. Again there is another low clock X2 coming up that we can't talk about just now, but fortunately you won't have to wait too long for details on that processor to emerge. Rather than focusing on increasing raw clock speed, AMD and Intel have both shifted to improving the features of their various chips. Dual core was the first step in that direction, and quad core (or maybe tri core) is a likely evolution at some point. 65nm processors are also on the horizon, and likely AMD will begin releasing the first such processors just beyond the range of the present roadmap - i.e. in late 2006 or early 2007. The process shrink should bring improved clock speeds as well as more cores per die. The increasing numbers of mobile parts are another indication of the changing goals. We may not be able to get much faster without spending significant effort, but we can try to focus on making the current designs more portable at the very least. We'll take a look at the Intel roadmap in the near future, but the trends are similar to what we see with AMD.
Facebook
Twitter
RSS
51 Comments
View All Comments
Zebo - Tuesday, July 26, 2005 - link
Hehe I hear ya this new format is confusing to say the least.. I much prefered the "flat" version. :)Anyway isolating CPU power today is faily trivial to do since all modern CPU's are fed by the 12V 4 or 8 pin dedicated auxiliary power connector. For specfic methodology please look here: http://www.lostcircuits.com/cpu/amd_venice/5.shtml">http://www.lostcircuits.com/cpu/amd_venice/5.shtml
JarredWalton - Saturday, August 6, 2005 - link
SO the 4-pin ATX12V is *only* for the CPU and the CPU doesn't draw power from anything else? Damn, learn something new each day! I guess a clamp around the two +12V lines would be able to measure the current with moderate accuracy.KristopherKubicki - Sunday, July 24, 2005 - link
Zebo: We have one coming up... And yes; Turion can't compare to Pentium M per watt.Pentium M isn't worthwhile outside of mobility; and no one argues it's an awesome mobility chip. Turion is nice, but AMD has an incredibly long way to go to catch up in the mobile sector. Expect an AnandTech review in the near future, although it's going to have a desktop focus.
Kristopher
Zebo - Sunday, July 24, 2005 - link
"AMD has been trailing in the mobile performance per Watt competition ever since Intel launched the Pentium M. "How do you know? You refuse to do a turion notbook review. Actually it's been 6 months since you've done a moblie review what's up with that? I sent anand an email, never heard back from him.
Anyway lets look at someone who HAS done the legwork;) :P Albeit not anandtech high quality standards.. just a few synthetics and batt life similarly equiped.
http://asia.cnet.com/reviews/hardware/notebooks/0,...
I know Intel fans think the Pentium M is the second coming of christ but the numbers simply don't bear that out.. Turion, at least in acer form, is very competitive performance and battery wise with the Pentium M, indescerable really, and it's usually cheaper.
Like this sweet little MSI jobber... Purdy and only $700
http://www.hardwarezone.com/articles/print.php?cid...
.. it's about time for Anandtech to clear up the noise with a real review don't you think?....
JarredWalton - Saturday, July 23, 2005 - link
20 - The current is reported from the AMD specs listed in the roadmap. While there will certainly be parts that run at lower power output than the maximum we've listed, the highest performing parts will likely approach those values.Maybe someone with a better knowledge of electronics can verify this, but just because the socket can supply 80 Amps on current 939 boards doesn't mean the CPU has to use all 80 Amps, right? It's like a water pipe going to the house: it might provide enough water pressure to run all the sinks at once, but the individual sink may never use that much water.
Anyway, 80 A * 1.4V = 112W. Revision E 939 is as follows: X2 chips have a TDP of 110W, meaning they can use everything the socket provides. FX-939 can use 104W, and the single core chips are 89W. Revision C/D used higher voltages but lower Amps and had FX-55 at 104W, FX-53 and other single cores as 89W. 60A * 1.7V = 102W, roughly the value listed for FX-55. The 1.5V chips were 90W in comparison.
Finally, we have M2 parts slated to use 95A. 95A * 1.3V = 123.5W, indicating that the FX will use all the Amps available. X2 will remain at 110W, so it will either use less of the available Amperage or else it will run at lower voltages. Single core will be 104W, slightly lower than X2. These are all maximums, however, so the mid-range parts probably won't be any worse than current parts; it's AMD building for future parts - they have to make sure that all motherboards can supply the power required by the top chips.
Jeff7181 - Monday, July 25, 2005 - link
Thanks for the response Jarred. Your response is accurate, but that doesn't explain this quote from the article.That's impossible unless the revision E San Diego's run on 1.125 volts. I understand 80A and 90W are "worst case scenario's." Still, it can't have a maximum amperage of 80 AND a maximum wattage of 90 because that means the voltage will NEVER be over 1.125 volts.
aldamon - Saturday, July 23, 2005 - link
With regards to your comments on the Newark 4000+:"As virtually any socket 754 board should have no trouble supporting these mobile variants, 754 owners looking to upgrade for additional CPU performance might be interested in checking out these parts."
My 8KDA3J won't support Newarks and as far as I can tell it's just a BIOS limitation. The 8KDA3+ is in the same predicament. It would be nice if Epox would make a effort to support ALL 754 CPUs for our older S754 boards. They're being stubborn so far.
PrinceGaz - Saturday, July 23, 2005 - link
#19 Jarred- you make some good points on whether the 4400+ is really a better buy than the 4200+. Yes I do intend to overclock as I'll be pairing whichever I go for up with a DFI mobo, and a Thermalright XP-90C heatsink (and suitable fan) which I've heard is one of the best coolers, and seeing what I can get out of it.The upcoming 3800+ is tempting as it should be a good bit cheaper, but I'm concerned that what they'll be are all the speed-binned rejects of both the Manchester and Toledo cored X2's that had very little headroom. The Toledo parts would have half the cache disabled of course, providing a further means for AMD to offload rejects with a fault in part of the cache (similar to what they did with .
I think I'll wait and see rather than possibly spend a small fortune on a 4400+ that overclocks no better than a part little more than half the price. Another month hardly matters as I'm not desperate, but it does seem a bit of a waste running my two new 1GB PC3200 CAS2 sticks on an old KT266A mobo at 138 FSB :)
Jeff7181 - Saturday, July 23, 2005 - link
Kristopher Kubicki & Jarred Walton... could you explain something to me?Watts = volts x amps
90 watt CPU running on 1.4 volts = 65 amps (90/1.4=~65)... how'd you get 80?????
JarredWalton - Friday, July 22, 2005 - link
13 - PrinceGaz, I realize that having more cache can be helpful, but in general is the $100 price increase worth the extra 4% performance for the X2 4400+? More importantly (if you're willing to overclock), it's probably $250 more than the X2 CPU that was mentioned above. and still only slightly faster. If I could get my hands on them, I'd like to try overclocking the 4200+ and 4400+. My instinct tells me the extra cache may reduce overclocking headroom a bit, making the two basically equal in performance.18 - I don't think DDR1 will dry up that quickly, so Q2'06 seems reasonable for DDR2. The 65nm parts from Intel are going to be the interesting competition. Pressler/Cedar Mill aren't too special, but Conroe/Merom could present a tought matchup for K8+DDR2. Unless K9 is coming out sooner rather than later, the new architecture from Intel may regain the performance crown for a while. Still, competition is good for us, so whatever happens happens. 200 MHz per quarter is going to be unlikely for a while, though. Some are saying we'll stay in the 2 to 4 GHz range for many many years and just add more cores.
As for DDR2/3 and FB-DIMM, while the base technology may be similar, I'd be surprised if the DIMMs are interchangeable. FB-DIMM is really targeting servers/workstations, where the current 2 DIMM per channel maximum memory is extremely limiting. It won't be quite as fast, but it should allow for 4 DIMMs per channel at least, and possibly more (?). Like registered vs. unbuffered DIMMs, the boards/chipsets/CPUs will either require FB-DIMM or not support it at all I think.