Wouldn't a Pentium 4 with a 533MHz FSB and dual-channel DDR266 be a perfect combination?

[Random Variable]

There wouldn't be any wasted bandwidth, right?

 

[Rand]

The peak bandwidth of a dual channel PC2100 DDR solution would match up perfectly with the bandwidth offered through the P4's 133/533MHz FSB, that said I wouldnt quite think of it as the 'perfect' solution to the P4/533.

It should be capable of providing some very nice competition against dual channel PC1033 DRDRAM though, and I suspect benchmarks would flip flop back and forth depending upon the requirements of the individual application.
Of course that could depend greaty upon who produced the chipset, Intel's DDR memory controller is certainly impressive on paper and has some nice specs but could still use some definite improvements IMHO before it's ready to match up to DRDRAM.

Dual channel PC2100 would likely cost slightly more then dual channel PC1033, but it should be low enough that it could be pushed into the mainstream

 

[MadRat]

The dual-channel DDR chipset for XEON isn't helping really any more than the single-channel chipset. It may not be the solution that people expect.

 

[Rand]

Originally posted by: MadRat
The dual-channel DDR chipset for XEON isn't helping really any more than the single-channel chipset. It may not be the solution that people expect.

Well, with the higher internal latency timings relative to the single channel I845 memory controller, and the fact that it's only running at PC1600 one shouldnt really be surprised. Internally it's different enough from the I845's memory controller that you can't really compare the two, and given the higher internal latency, the performance reduction from using ECC, and also the slightly higher latency as a result of running at PC1600 rather then PC2100 I'm not at all surprised it's failed to match up to the I860... still most benchmarks are putting it reasonably close.

Close enough that I suspect a dual channel dirivative of the present I845 would offer reasonable competition to the dual channel PC1033, though IMHO Intel at least still needs to improve the implementation of the DDR memory controller before it'll be quite up to par with the dual channel PC1033.