Replies to post #7450 on Advanced Micro Devices Inc (AMD)

[Haddock]

jhalada,yb I remember there was an Opteron & Athlon 64 Assymetric Low Cost Desk-top on slide 47 of the Platform Conference presentation. I couldn't find it on AMD's web site but google showed this:

Interesting design, but difficult to explain to customers. Would AMD validate the A64 for a dual design like this? The Opteron would be a 2xx.

Would make a nice mobo with nForce3.

For those that can't be bothered to download a ppt file:

Mem
-----------
/
Athlon
/
Opteron---___________mem
/
IO

[jhalada]

thanks for the reminder. That would make a viable system, but would make sense only if prices of Opteron 2xx and Athlon64 were vastly different.

On the memory types, I am hoping that one of the hardware sites will give write an article explaining what's possible. Is it possible to have unbuffered memory on Opteron, registered on Athlon64?

Joe

[sgolds]

chr_p, fasinating find, that assymetric Opteron/A64 design! It does get over the problem we mentioned yesterday that you can not make a practical configuration with two A64s. In that case, the lone aHT connection from the A64 goes into one of the connections on the Opteron, and the Opteron handles all the I/O.

Technically feasible, but silly, IMHO. First, it would require the OS to support distributing threads based on the characteristics of the microprocessor. I'm not sure how many applications can be written this way. Next, applications tend to do most of their hard work in the background tasks (because the foreground task is interacting with the users). (Note: Some games are an exception to this rule.) For most things, the CPU-hungry background task is being delegated to the wrong CPU!

Your question: Why use regular DDR on the Opteron? Saves cost, improves performance over buffered DDR. I'm not sure you can actually populate a processor with 8GB of regular DDR right now, but this slide projects such a configuation. Usually you use buffered DRAM to get this size configuration.

I see this diagram as a bit of 'blue skying' - showing what is possible, not what is feasible.

I think I will send Peter an email - his address is on the first slide. Maybe he can shed some more light on it.