<Intel's management appeared very confused about the bandwidth issue when they jumped on board the Rambus bandwagon last year, and I made numerous posts back then on the Intel and AMD threads discussing this.>
And like I said before, Scumbria, if you have a solution to the latency problem, we'd all like to hear it. None of the alternatives to DRDRAM reduces latency by any significant amount.
I'll bet even Intel's management is aware of the latency issue, that there isn't as much "confusion" over the problem as you mention. There are several other reasons why Intel is going gung-ho with DRDRAM, such as low pin count, more efficient use of bandwidth (thereby reducing average latency in a loaded system), viability for Timna platform, performance for the high-end, etc.
By the way, why is your company going with DRDRAM for your graphics design? In other words, what are the main advantages with going to DRDRAM over the alternatives, in specific terms?
Tenchusatsu |
