KZ--After I supported your previous post, please don't go unbalanced on me. You said [bold added],
An apples-to-apples comparison would be to compare it to PC133 with a chip set with 133MHz FSB. And
before those apps demanding bandwidth appear DDR200 and DDR266 will appear in full production and provide
more bandwidth (or at least equal bandwidth to Rambus). Following that DDR-II going above 400 MHz will
appear. All of this will occur in the not distant future as VIA perfects their chip set and other vendors come on
stream. Actually right now is the best time for Rambus performance and it will get worse relative in the future.
Within 6 months I predict that "Rambus performance penalty" will enter the language just like Rambus die size
penalty has. At that point it will be obvious to everyone that for PCs Rambus has higher cost, lower performance,
and lower reliability.
The question in each of your points, of course, is, WHEN?!!
I don't mean to be too flippant on this, but we don't know when any of these will be successfully commercialized. We can at least agree on that, can't we? A race is in progress, both among the competing approaches and probably a terminal date by which none of the presently known competitors will be useful.
At least implicit in your post is your acknowledgement that PC133 is a dead end. Thank you for that.
As generations of DDR evolve, don't you expect that DRDR will evolve, also?
There is a huge question surrounding when rdram will work with the 820 chipset. A protracted delay will substantially erode the opportunity to dominate desktops. There is no question that it is the method of choice in current graphics-intensive hardware.
Best,
JS |
