After we filed Peter Sherriff's earlier story about Coppermine being delayed because of marchitecture rather than architecture issues, an Intel engineer in Israel has emailed us to put the record straight. (Story: Intel Cumine delay caused by fear of AMD) At the same time, she also commented on our stories about AMD's process in Dresden. Our source, who wishes to stay anonymous because she wants to keep her job, but who is nevertheless, extremely reliable, said: "I can assure you that Coppermine's delay had nothing to do with marketing. For once, Intel's PR is telling the truth. There was a RASH of unexpected problems with it, mostly due to last minute problems hitting the frequency target." She said that Dixons early success was because it failed to fully exploit the .18 micron P858 process. Coppermine, which does (or perhaps should) fully exploit P858, showed a number of problems with both process and design. Coppermine will launch at 667MHz because of the holdover P856 version, she said. Intel will take a rather standoffish position towards AMD's Athlon, she said. There will be no official reaction while architects, engineers and designers "fab out" (make) faster and faster Cumines until Willamette emerges. And the engineer further claimed that there is nothing in AMD's Fab 30 that Intel hasn't already tried. However, she added, Intel doesn't promise things it can't produce... Hmmm... Flip chip technology is in place all over the place at Intel, she said, since the Pentium II/350. (We do remember that our old friends at IBM Microelectronics were talking about this flip chip technology in 1992. Where is Mr La Rosa these days? Is Via gonna buy IBM too? This chip stuff is dead tricky, init?) ®