Very strange to post a reading list from a course given in Summer 2002.
I took this same class at UCF more recently with Dr. Heinrich, who did an amazing job. I don't think his course materials are public, though he used Modern Processor Design[1].
WRT the Cray article, I think it would be a struggle to find a better ten page "primary-ish" source than the original ACM article for the Cray-I. Bitsavers (and mirrors) has the original hardware ref manual but that's in excess of two hundred pages. The original marketing brochure (see bitsavers again) is basically the linked ACM article with bigger graphics and is 15 pages long, better off just reading the shorter ACM article. Its relevant today because as a marketing tool back then it was the first heavily pipelined vector processor for a lot of otherwise experienced folks which in 2013 makes it an ideal intro to the topic for comp sci students.
I'm having severe trouble at this time of day thinking of anything new in processor design over the past decade to add to a list. New implementations of very old ideas, bigger implementations of very old ideas, but not able to think of any actually new ideas. I'm talking about a decade of substantial engineering-type achievements vs no basic science-type discoveries.
_Modern Processor Design_ (Shen and Lepasti). Lots of material on superscalar processors, with some nice examples from existing work.
_The Pentium Chronicles_ (Robert Colwell). Much less technical, but a lot about the social nature of getting a significant CPU design done.