They're comparing to an in-order CPU. Given that most CPUs are out-of-order (at least of the non-embedded variety, and GC is less used in such applications anyway), it would be better and more intellectually honest to actually compare to a typical CPU that performs GC. They kind of address this in the paper but only in a short aside: "Note that previous research [1] showed that out-of-order CPUs, while moderately faster, are not the best trade-off point for GC (a result we confirmed in preliminary simulations)." So they don't quantify what any of this means.
I think it's an interesting idea, but it doesn't bode well when they seemingly choose the wrong target for comparison and hand-wave away the difference as insignificant.
I'm pretty sure that is the first implementation of Lisp. Earlier "versions" of Lisp were hand-compiled, because it was considered too difficult to produce a compiler. Then Steve Russell realized that McCarthy's "eval" function could be implemented in assembly language (which apparently hadn't occurred to McCarthy, he considered it merely theoretical) and produced the first Lisp interpreter. Note that the manual is from March 1960 and McCarthy's first paper on Lisp was published in April 1960, so nobody outside his circle at MIT would have known about Lisp at the time.
The manual also states that it's for "a version of Lisp being prepared for the IBM 709", implying the manual was written while the interpreter was still being developed, and all references I've found state that the IBM 709 version was the first practical implementation.
I think it's an interesting idea, but it doesn't bode well when they seemingly choose the wrong target for comparison and hand-wave away the difference as insignificant.