NForce3 250 - looks promising

http://www20.tomshardware.com/mothe...b_passes_the_athlon64_performance_jump_hurdle

So much for socket 939 to be the long lasting platform, when DDR comes out in '05, they'll need new motherboards again. Oh well, this has a lot of goodies in it.

Shocking...

Huh... Um would the CPU socket need to change to accomodate DDR2? I thought there would just be new 939 boards coming out at around that time with DDR2 ram sockets.

/IMOG goes to read article

EDIT: LINKY no GOODY!

Error404 - Document Not Found!

Sorry! The Requested Page was not Found!!

Dear Visitor. The page you requested couldn't be found - If you followed a link from another Website please inform their Webmaster. If you happen to get this message while browsing Tom's Hardware Guide please inform the Webmaster of Tom's Hardware Guide.

Please check the URL in the address field of your browser. All pages end with .html, maybe it is .htm. You might want to use the Search Engine to find what you are looking for.

You can also browse Tom's Guides (see box below). They list articles according to their topic. Or you might want to check the Historical Archive. Here you can find all articles in a chronological order. The Keyword Index provides an alphabetical listing of products, product groups and companies.

Click to expand...

This might be the right working link:

http://www20.tomshardware.com/motherboard/20040420/index.html

While AMD wants to downgrade socket 754 almost to a pure 32 bit platform, hopes for success in 2005 are based on the Toledo 90 nm chip and socket 939. If the processor were to get a DDR2 memory interface by then, we would once again expect a new processor socket to be launched. Since DDR2-DIMMs have 240 rather than 184 contacts, operation of a DDR2 Athlon64 in socket 939 is not possible with conventional 32 bit DDR2 DIMMs. One possible solution in that case would be 64 bit modules.

Click to expand...

Huh? Dunno if it is just the wording here or what... But it seems like the author is trying to say something that he doesn't understand. Whatever he is trying to say, I don't understand it.

the author DEFINATLY sounds like he's trying to solve a problem that he doesn't know anything about...

Go figure, it's from Tomshardware.......You will find that over there.

I'm curious now...

This is what I typed in this thread, also discussing Nforce3-250 a bit:

Mention of NF3-250 here... Might want to take a look. DDR2 will require another new CPU socket, other than 939? I thought the RAM sockets were all that would be different (like a socket A mobo that takes SDR 168 pin RAM, or a socket A mobo that takes DDR 184 pin RAM)?

Click to expand...

I have no idea why DDR2 would require a different CPU socket... I would really like it if someone could comment who understands enough to know why it would or wouldn't.

DDR2 and DDR memory modules are different in terms of pin count, voltage, signal timing, signal termination, ..., as far as the memory module and CPU/memory controller interface is concerned. Definitely a new motherboard layout would be required, but this may not be sufficient for the DDR2 change over.

As such difference, the socket 939 pin layout may or may not be able to interface with the DDR2 memory module, given the constraint imposed by the layers of motherboard interconnect, signal to noise consideration, ...., and more engineering details that have to be addressed. If it can be resolved, who would want to have another socket.

AFAIK, I don't think there is a yes or no answer yet.

Of course, if everything could be perfectly planned ahead for few years, one might be able to foresee and came up with a socket pin layout that can accomodate for DDR memory module, DDR2 memory module, 64-bit memory bus, 128-bit dual channel memory bus, ...., then we would not have to face with this famous 940, 754, 939 socket migration problem and uncertainty, ....

The first 940 Athlon 64, "Opteron that is not an Opteron" as Stroligo calls it is an indication of what AMD planning is.

Current assumptions:
1. Before Toledo there will be San Diego.
2. San Diego will not use DDR2.

Toledo may not be compatible with initial 939 mobos just like Thoroughbred wasn't with some Palomino boards just like Barton wasn't with even more Palomino/Thoroughbred boards even though they were ALL socket A boards.


P.S. hitechjb1, I don't think your popup PM message notification is on.

I have read a few times that AMD have stated that WHEN they do introduce DDR2 support in their a64 memory controllers it will be in a 939 pin layout - so it would seem they think it is entirely possible.
One such example here.http://www.xbitlabs.com/news/memory/display/20040326063601.html

If I read the various articles correctly, DDR2 won't happen at least until late 2005. It may stay with socket 939. But that is not a final official statement from AMD. I think there must still be some engineering and roadmap issues have to be ironed out.

A simple look at multipliers:
http://www.c627627.com/AMD/AthlonXP/
shows there's room to scale the 64 with San Diego but that there is no room to do so with Toledo.

AMD will have to either switch to DDR2 with Toledo or switch to one of the higher rated DDR1's unofficially available even today:
http://www.c627627.com/RAM/

From what I have seen - the DDR2 is expected to be introduced with the Toledo - however that is entirely down to whether DDR2 meets the 2 stipulations that AMD mentioned in march regarding price and performance
http://www.vr-zone.com/?i=632&s=1
http://www.anandtech.com/showdoc.html?i=2006

"As for the delay with DDR2 support, AMD is going to make up for it no sooner than the second half of the next year. So far, they have limited themselves to using memory modules on DDR500 chips."
http://www.xbitlabs.com/articles/editorial/display/apr2004_2.html

Will they use DDR500 with San Diego OC D, are you ready to call them on that + if they did:
[250] x 11 = 2750
[250] x 11.5 = 2875
[250] x 12 = 3000


The 939 ClawHammer preceeding San Diego will probably be a 2.6 or a 2.8 part, so what do you think?