- Sep 20, 2002
- 1,382
- 0
- 0
Edit: Link to article
I spent some time reading over the article, which I think was very good. It explains how SDRAM works quite clearly. I also have some questions concerning the future of system memory. In DDR2 memory, 4 SDR inputs are multiplexed onto 1 DDR output at 2x the frequency. Will future memory work by multiplexing more than 4 inputs, either 8 or 16 at 4x or 8x the frequency, or is there another way to increase the performance of SDRAM. There is the ability to increase the actual clock of the memory cells, but that method is trivial (obvious.) I remember Via(?) was pursuing QDR memory modules back in '02, is that likely to be adopted? I imagine that the more SDRAM evolves, the more it will resemble RDRAM.
On another line of thought, what kind of research are we doing to find a replacement to DRAM? That's enough from me. Thanks. Peace.
I spent some time reading over the article, which I think was very good. It explains how SDRAM works quite clearly. I also have some questions concerning the future of system memory. In DDR2 memory, 4 SDR inputs are multiplexed onto 1 DDR output at 2x the frequency. Will future memory work by multiplexing more than 4 inputs, either 8 or 16 at 4x or 8x the frequency, or is there another way to increase the performance of SDRAM. There is the ability to increase the actual clock of the memory cells, but that method is trivial (obvious.) I remember Via(?) was pursuing QDR memory modules back in '02, is that likely to be adopted? I imagine that the more SDRAM evolves, the more it will resemble RDRAM.
On another line of thought, what kind of research are we doing to find a replacement to DRAM? That's enough from me. Thanks. Peace.