Jeffrey and SMALL FRY, re: "Intel sticks with aluminum at 0.18 microns."
Intel went public about six months ago with the position that they would stay with Aluminum vs. going over to copper for the 0.18 micron geometry process . IBM and then Motorola made some splashes back then about Cu, saying its lower resistivity than Al would, with the very high logic density and , therefore, power in the latest chips, start to make a difference in power dissipation and speed. Also, they have the Cu process "ready". Intel countered that, for now (0.18) it is more important to continue to improve on-chip transistor speed characteristics and lower k dielectric. From the Cyrix thread, process development consultant Yousef:
Intel is running at 900mhz on
their SRAM process development test chip. The corresponding
4Mb test
chip in the .25um process I believe ran at about 350-400mhz. This
just
indicates that there is some more "Mhz headroom" on their new
.18um process.
Also, Intel is using a power supply voltage of a scant 1.5 volts getting that speed. This, and Yousef's comment says it all.
In the meanwhile, IBM, #1 proponent of Cu, has yet to announce a chip even at 450 MHz, whereas Intel will be there with 500, 533 and up in 1H99 (Katmai et al).
Also from the Cyrix thread, some yahoo named Tony:
Copper, we don't need no steenking copper!
However, Intel has announced the will go to copper for 0.13 micron, in about 2001, don't know if 100%, or not.
Tony |
